The Holocene–Anthropocene Transition in Lakes of Western Spitsbergen, Svalbard (Norwegian High Arctic): Climate Change and Nitrogen Deposition

J Paleolimnol DOI 10.1007/s10933-009-9338-3

ORIGINAL PAPER

The Holocene–Anthropocene transition in lakes of western Spitsbergen, Svalbard (Norwegian High Arctic): climate change and nitrogen deposition

Soﬁa U. Holmgren Æ Christian Bigler Æ O´ lafur Ingo´lfsson Æ Alexander P. Wolfe

Received: 4 April 2008 / Accepted: 22 April 2009 Ó Springer Science+Business Media B.V. 2009

Abstract Lake sediments from four small lakes on probably related to climate warming after the Little western Spitsbergen (Svalbard Archipelago, Norwe- Ice Age. However, the most pronounced changes in gian High Arctic) preserve biostratigraphic and diatom assemblages seem to have occurred in the last isotopic evidence for a complex suite of twentieth few decades. At the same time, nitrogen stable century environmental changes. At Lake Skardtjørna isotopes in sediment organic matter in two of the and Lake Tjørnskardet on Nordenskio¨ldkysten, there lakes became progressively depleted by *2%, which is a marked diatom ﬂoristic change coupled to is consistent with diffuse atmospheric inputs from increased diatom concentrations beginning around anthropogenic sources and attendant fertilization. 1920. At Lake Istjørna and Lake Istjørnelva, 25 km These data suggest that climate change and nitrogen southwest of Longyearbyen, both diatom total valve deposition may be acting together in driving these and chrysophyte stomatocyst concentrations have lakes towards new ecological states that are unique in increased dramatically since the beginning of the the context of the Holocene. 1900s. The early twentieth century changes are Keywords Anthropocene Arctic Diatoms Nitrogen Paleolimnology Spitsbergen S. U. Holmgren (&) Department of Earth Sciences, University of Gothenburg, 405 30 Gothenburg, Sweden e-mail: soﬁ[email protected] Introduction

S. U. Holmgren O´ . Ingo´lfsson Lake sediments from many regions of the Arctic The University Centre in Svalbard, P.O. Box 156, chronicle pronounced environmental changes in the 9171 Longyearbyen, Norway late nineteenth and twentieth centuries (Douglas et al. C. Bigler 1994; Gajewski et al. 1997; Wolfe and Perren 2001; Department of Ecology and Environmental Science, Sorvari et al. 2002; Birks et al. 2004; Smol et al. Umea˚ University, 901 87 Umea˚, Sweden 2005). In general, the biological shifts reported in O´ . Ingo´lfsson these studies are consistent with hypothesized eco- Department of Earth Sciences, University of Iceland, logical responses to recent climate warming, includ- 101 Reykjavik, Iceland ing longer growing seasons for lake biota as a consequence of reduced summer lake ice cover. The A. P. Wolfe Department of Earth and Atmospheric Sciences, current warming trend in the Arctic appears to have University of Alberta, Edmonton, AB T6G 2E3, Canada started in the mid nineteenth century, by the end of 123 J Paleolimnol the Little Ice Age, although considerable regional species responses to N and N ? P amendments are differences exist (Overpeck et al. 1997). Anthropo- pervasive (Gordon et al. 2001). genic influences are believed to have caused an Here, we consider diatom and sediment geochem- acceleration of this warming trend in recent decades, ical records from four small lakes on western as witnessed in many, but not all, areas of the Arctic Spitsbergen that reveal striking changes in sediments (Serreze et al. 2000). Although the Eurasian sector of deposited in the twentieth century. We compare these the Arctic appears to be currently warming at a rate data with the early to mid-Holocene variability on the order of 0.5°C per decade (Comiso 2003), it recorded in a long core retrieved from one of the remains uncertain whether current summer tempera- lakes as well as with available climatic and glacio- tures are higher than those of the early Holocene, chemical time-series in order to explore the causative when summer insolation was considerably higher factors driving the recent observed stratigraphic than present (Berger and Loutre 1991), and glacier changes in the lakes. Our investigation builds upon retreat was extensive (Koerner and Fisher 2002). previous studies of the recent paleolimnology of Cirque glaciers are thought to have retreated consid- Svalbard (Birks et al. 2004) that concluded that no erably or even disappeared from many catchments in clear responses to atmospheric deposition could be western Svalbard, only to re-form during the late detected, despite recent changes in several proxies. Holocene (Svendsen and Mangerud 1992, 1997). The unprecedented character of biological shifts in many Arctic lakes during the twentieth century (Jones and Study sites Birks 2004; Smol et al. 2005) strongly implies that current environmental conditions are fundamentally The study sites (Fig. 1) are four small non-glacial lakes different, in some regard, from those of the past, in Nordenskio¨ldland, western Spitsbergen; two close to which raises the question as to whether climate the coastal strand flat of western Spitsbergen, and two warming is the only factor involved in these changes. south of Colesdalen further east. The two western lakes Analyses of ice cores from around the Arctic are situated in the Tjørnskardet Pass. They occupy (Mayewski et al. 1990; Goto-Azuma and Koerner catchments underlain by metamorphic rocks of Late 2001; Kekonen et al. 2002; Isaksson et al. 2003) Precambrium (Hecla Hoek) age (Hjelle et al. 1986). 0 0 consistently reveal increased NO3 and SO4 concen- The lower Lake Skardtjørna (77857 N13849 E, trations in the late twentieth century, indicating that 61 m asl) is dammed behind a raised beach that forms polluted air is effectively delivered to the Arctic from the postglacial marine limit at 65 m asl (Landvik et al. mid-latitude Eurasian and North American sources. 1987), whereas the second lake, Lake Tjørnskardet However, resultant rates of anthropogenic nitrogen (778580N138520E, 120 m asl) occupies the col of the deposition have been considered insufficient to cause pass. The two lakes further to the east, Lake Istjørna ecological changes in high-latitude terrestrial com- (788020N158080E, 235 m asl) and Lake Istjørnelva munities (AMAP 1997, 1998). The sensitivity of (788000N158140E, 210 m asl) are situated in Istjørnd- Arctic lakes to nitrogen deposition has not been alen, south of the valley Colesdalen. Their catchments tested explicitly, but given responses observed in are underlain by Tertiary shale and sandstones (Major other low-nutrient lake systems (Sickman et al. 2003; and Nagy 1972). The area in general shows few traces Wolfe et al. 2003; Bergstro¨m and Jansson 2006), it of glaciation, with only a sparse till cover. Extensive has been predicted to be high. Alterations on the postglacial colluvium (talus, rock glaciers) occupies order of 10 kg ha-1 year-1 are considered sufficient the valley sides. Vegetation in the area is limited, in to induce changes in Arctic tundra (Gordon et al. part due to pronounced solifluction. In general, the area 2001). Furthermore, the possibility exists that nitro- lies within the Dryas octopetala zone of Brattbakk gen availability will increase as climate warms, as the (1986). result of accelerated soil mineralization, quite inde- Lake Skardtjørna is kidney-shaped with a shallow pendently of atmospheric deposition rates (Hobbie basin that dries out partly during summer, in the et al. 1999). On Svalbard, it has been suggested that southwest, and a deeper basin in the middle of the the critical load of N deposition for terrestrial elongated lake, to the east. The lake has an area of ecosystems is lower than for mid-latitudes, and that 0.097 km2 and a maximum depth of 7.3 m. The 123 J Paleolimnol

Fig. 1 Map of study area showing locations of Lake Skardtjørna, Lake Tjørnskardet, Lake Istjørna and Lake Istjørnelva

catchment area is 1.24 km2. A rock glacier occupies airport is -5.7°C as an average for the last 30 years the southeast side of the lake. Lake Tjørnskardet is (1973–2002) or -5.1°C as the 11-year running mean located about 1.5 km from Lake Skardtjørna. It is temperature (1992–2002), with July temperatures narrow, elongated and shallower than Lake Skardt- of 6.3°C (average 1973–2002). Mean annual jørna, with a surface area of 0.081 km2 and a maximum precipitation is 192.5 mm on a 30-year average depth of 2.45 m. The catchment area is 2.2 km2. Lake (1973–2002), and 197.9 mm on an 11-year running Istjørna is oval-shaped and the largest and deepest lake mean (1992–2002). in this study, with a lake area of 0.37 km2 and a maximum depth of 29 m. The catchment area is 1.76 km2. Lake Istjørnelva is a small, elongated lake, Materials and methods situated about 4.5 km southeast of Lake Istjørna. It is the smallest of the lakes in this study, with a surface area Sediment coring and chronology of 0.047 km2 and a maximum depth of 4.5 m. It has no inflow or outflow. The catchment area is 0.33 km2. Short cores from four lakes were obtained in April Additional limnological data are provided in 2001 with a modified Kajak-Brinkhurst gravity Table 1. Mean annual temperature at Longyearbyen device that preserves the sediment-water interface 123 J Paleolimnol

Table 1 Water chemistry of the four study lakes, and major ion chemistry from rain from Linne´dalen Water chemistry Lake water samples Precipitation Linne´vatnet (mg/l) L. Skardtjørna L. Tjørnskardet L. Istjørna L. Istjørnelva LP001 LP002 LP003 LP004

- NO3 (mg/l) 0.40 0.30 DL 0.19 0.21 1.27 0.16 1.56 ? NH4 (mg/l) 0.23 DL 0.18 0.01 2- SO4 (mg/l) 19.79 4.75 18.06 66.19 0.84 2.27 1.48 2.08 3- PO4 (mg/l) DL DL DL DL DL DL DL DL - HCO3 (mg/l) 133.85 67.96 4.51 9.15 F- (mg/l) 0.13 0.12 0.18 0.12 Cl- (mg/l) 10.67 7.69 5.35 5.81 0.64 1.58 1.86 0.09 Ca (mg/l) 31.00 16.50 2.70 11.80 K (mg/l) 1.26 0.28 0.63 1.19 Mg (mg/l) 11.40 5.06 1.68 5.27 Na (mg/l) 5.16 3.28 5.59 12.40 P (mg/l) DL DL \0.1 \0.1 S (mg/l) 6.63 1.75 5.74 23 Si (mg/l) 0.92 0.35 0.62 0.35 pH 7.39 7.28 6.29 6.38 Conductivity (lS/cm) 246.60 128.90 65.10 171.90 Lake area (km2) 0.097 0.081 0.370 0.047 Lake depth (m) 7.30 2.45 29.00 4.50 Catchment area (km2) 1.24 2.20 1.76 0.33 DL below detection limit intact (Glew 1989). The cores were extruded in the calibrated to calendar years using IntCal98 (Stuiver ﬁeld into continuous 0.5-cm increments throughout, et al. 1998). except for the core from Lake Skardtjørna, which was subsampled into 1-cm increments below 15 cm Diatom and chrysophyte stomatocyst analysis depth. The Lake Skardtjørna core was 38 cm long, Lake Tjørnskardet 19 cm, Lake Istjørna 34.5 cm and Subsamples of 200-mg, freeze-dried sediment were

Lake Istjørnelva 19.5 cm long. Samples were trans- oxidized with 15% H2O2 for 24 h, then 30% H2O2 for a ferred to plastic bags, transported to the labora- minimum of 24 h, and finally heated at 90°C for tory, and freeze-dried for permanent archiving. For several hours. The slurries were then centrifuged and 210 sediment geochronology, Pb (t1/2 = 22.3 years) rinsed with deionised water, then diluted to 10 ml. A activities were inferred from measurement of its known quantity of markers (Eucalyptus pollen) was grand-daughter 210Po using alpha spectroscopy, and added to 200-ll aliquots of the digested slurries to ages were determined by applying the constant rate calculate diatom concentrations (Battarbee and Kneen of supply (CRS) model (Appleby and Oldfield 1978). 1982; Wolfe 1997). These secondary slurries were A continuous depth-age model was developed by diluted again, mixed, pipetted onto coverslips, dried at fitting polynomial splines. For Skardtjørna geochro- room temperature, and mounted in NaphraxÒ medium nology 137Cs was also measured. To assess natural (refractive index = 1.65). At least 400 diatom valves Holocene variability of sediment components, a 5.5- per sample were counted except where the concentra- m sequence was obtained with a Russian peat corer tions of diatoms were very low, i.e. where there were (Glew et al. 2001) from Lake Skardtjørna. The almost no diatoms at all. Diatom identification chronology of the long core was established with 7 followed largely Foged (1964), Krammer and Lange- AMS 14C dates (Table 2) on terrestrial plant macro- Bertalot (2001) and van de Vijver et al. (1999). Diatom fossils (mostly leaves of Salix polaris), which were results are expressed as relative abundances for each 123 J Paleolimnol

Table 2 Radiocarbon dates from Lake Skardtjørna Sample depth (m) Lab. no. Material analysed 14C age Cal years BP

0.165 LuA5242 Salix polaris leaves 50 ± 85 50 1.1995 LuA5247 Salix polaris leaves 1,205 ± 85 1,210 2.2165 LuA5244 Salix polaris leaves 2,390 ± 90 2,410 3.181 LuA5246 Salix polaris leaves 3,565 ± 95 3,890 4.098 LuA5241 Salix polaris leaves 4,280 ± 90 4,900 5.225 LuA5245 Salix polaris ? Dryas octopetala leaves 6,775 ± 95 7,660 5.475 LuA5243 Salix polaris leaves 7,365 ± 90 8,230 All dates were performed on terrestrial plant macrofossils (leaves of Salix polaris), which were then calibrated to calendar years using IntCal98 (Stuiver et al. 1998) taxon, and as total concentrations (valves per g dry Sediment geochemistry sediment) estimated by the counts of introduced markers. Chrysophyte stomatocysts were found in For two of the lakes, Lake Skardtjørna and Lake small concentrations in all four lakes, but only Istjørna Tjørnskardet, sediment carbon and nitrogen contents and Istjørnelva had sufﬁciently high concentrations for were determined by pyrolysis in a Carlo-Erba 1500 counting. At least 300 cysts for Istjørna and 100 cysts CN analyzer. For determining organic carbon, dupli- for Istjørnelva were counted at each level and results cate samples were analyzed after pretreatment with are shown as absolute abundances of the total number 30% HCl to remove carbonates. The data are also of specimens (cysts per g dry sediment). expressed as C:N molar ratios to discern changes in Detrended canonical correspondence analysis sources of the organic matter in the lakes (Kaushal (DCCA) (ter Braak 1986) was used to obtain estimates and Binford 1999). of compositional turnover as beta-diversity, scaled in Measurements of the nitrogen stable isotopic ratio standard deviation (SD) units (ter Braak and Verdons- (d15N) were undertaken on sediment organic matter chot 1995). Sample ages based on 210Pb dating were to detect possible changes in nitrogen biogeochem- used as the only constraint in the DCCA to obtain istry (Talbot 2001). estimates of the total amount of compositional change Nitrogen isotopes were determined by isotope-ratio in the diatom record since AD 1850. The DCCA mass spectrometry (Finnigan Deltaplus XL, and are analyses were performed using CANOCO 4.0 (ter expressed as d15N(% relative to air) values as follows: 15 15 14 15 14 Braak and Smilauer 2002). d N = (( N/ Nsample)/( N/ Nstandard) - 1) 9 1000. For Lake Skardtjørna and Lake Tjørnskardet, diatom species richness was assessed by rarefaction analysis (Birks and Line 1992) to produce unbiased Results records of changes in diatom community structure in the lakes. Because the number of diatom species Sediment lithology and chronology encountered in a sample depends on the number of individuals counted (Birks and Line 1992), the The sediments of Lake Skardtjørna consist of lami- richness of diatoms for different samples is only nated, brownish-grey, slightly silty, clayey gyttja comparable if the counts are standardized. This can with darker grey laminations. Lake Tjørnskardet be done with the statistical method of rarefaction sediments consist of massive, olive-brown, slightly analysis (Birks and Line 1992). Rarefaction is a organic, silty clay with no apparent laminations. Lake numerical method for obtaining estimates of the Istjørna and Lake Istjørnelva sediment consist of number of diatom taxa expected in counts of a ﬁxed brownish-grey, slightly organic, silty clay with no number of diatom valves. In our study, the count size apparent laminations. chosen for standardization was set to 100 in both The 210Pb activities in both cores show strati- lakes. graphic reversals (i.e. up-core decreases in activity), 123 J Paleolimnol

Fig. 2 Pb-210 activity profiles and CRS age-depth model for Lake Skardtjørna. Shaded intervals indicate dilution of the results, fitted with polynomial splined curves for the four study 210Pb activity during episodes of accelerated sedimentation lakes. Cs-137 activity profile and associated age-depth model 2 especially marked in the Lake Skardtjørna sediments Diatom taxa in samples with less than 100 identified (Fig. 2). These reversals are most likely caused by specimens are not shown as percentages, but marked episodes of accelerated sedimentation rates, diluting with a dot that indicates ‘‘presence.’’ Between 25 and the 210Pb. The 210Pb/226Ra equilibrium, which corre- 43 samples in each lake were analysed. In two of the sponds to ca. 150 years of accumulation, occurs at ca. lakes (Lake Istjørna and Lake Istjørnelva) concentra- 17.5 cm depth in Lake Skardtjørna, 11.5 cm depth in tions were close to zero prior to 1900. In all lakes there Lake Tjørnskardet and at 5.75 and 6.75 cm depths in is a dominance of periphytic taxa. Lake Istjørna and Lake Istjørnelva, respectively. For Lake Skardtjørna, 137Cs was also measured, clearly showing the 1963 atomic bomb test maximum and Lake Skardtjørna peaks at 1954 and possibly at 1986. For comparison with the CRS age model, a 137Cs age-depth model At Lake Skardtjørna the diatom assemblage is domi- was made from four time markers, i.e. 1986, 1963, nated by Fragilaria pinnata throughout the whole core, 1954 and 1850, the depth at which the first unsup- with other species only making up a minor part of the ported 210Pb appears. We assumed a linear relation- assemblage (Fig. 3). In the lowermost samples ship between these points (Age model 2 in Fig. 2). (37–24 cm), F. pinnata is the only taxon recorded. From ca. 24–11 cm, the F. pinnata abundances Diatoms and chrysophyte stomatocysts gradually decrease upwards to around 80%, while other taxa such as Fragilaria pseudoconstruens, The diatom records are presented as diatom percent- Amphora, Navicula and Nitzschia appear, of which ages and as diatom total valve concentrations (Fig. 3). F. pseudoconstruens is the only species reaching 123 J Paleolimnol

Fig. 3 Relative frequencies of dominant diatom taxa in cores from the four lakes. Included are diatom total valve concentrations and chrysophyte stomatocyst concentrations. CRS 210Pb ages are indicated to the right

an abundance of 10% in some samples. Above ca. uppermost samples. The ﬂora now also includes 10 cm a marked shift in diatom assemblages occurs, as F. pseudoconstruens, Amphora inariensis, A. ovalis, the relative abundances of F. pinnata start decreas- A. libyca, Navicula, Nitzschia and small Achnanthes ing markedly, to abundances of \40% in the species. Dissolution is apparent, but mostly in the base 123 J Paleolimnol of the core. It is impossible to say to what extent and declining F. pinnata percentages. After the dissolution has affected the diatom total valve con- dilution episode, the concentration of non-Fragilaroid centrations. Because dissolution is only obvious in the taxa increases. Furthermore, during the episode of base of the core, it probably does not affect concen- increased sedimentation rate, Amphora ovalis occurs trations very much in the top half of the core. in its highest abundance throughout the whole core. In Lake Skardtjørna, there is an increase in total The F. pinnata concentrations increase immediately diatom valve concentrations, interrupted by a dilution after the dilution, probably as a consequence of episode with increased sedimentation rates at ca. reduced dilution, following the decrease in sedimen- 12–10 cm, i.e. at the end of the 1940s. Subsequently, tation rate. Thereafter, the concentration slowly lower and fairly constant diatom valve concentrations decreases towards the top. The uppermost ca. 4 cm, are recorded. However, the non-Fragilaroid valve after the late 1980s, seem to be affected by another concentrations show an increasing trend from ca. episode of increased sedimentation rate, but less 15 cm depth, corresponding to the early 1900s, which extreme than the previous event and without apparent is particularly pronounced above 3 cm, i.e. after dilution of the diatom concentrations. The uppermost *1990s. Increased sedimentation rate dilutes the *3 cm show the most pronounced increase in concentration of diatoms, but the episode in this case concentration of non-Fragilaroid taxa. Above ca. is also coupled to a ﬂoristic change in the diatom 3cm(*1990) a minor assemblage change occurs, assemblage. Lake Skardtjørna exhibits a very dra- with declining abundances of F. pinnata, increased matic change in diatom species composition at ca. values of Navicula cryptotenella, Nitzschia permin- 10 cm (*1950), with the introduction of new uta-frustulum complex, small Achnanthes spp., species, the distinct increase in species diversity, A. inariensis and Amphora ovalis. The diatom

Fig. 4 Paleolimnological and biogeochemical data. Diatom carbon and total nitrogen content (% of dry mass), C:N molar total valve concentrations and non-Fragilaria valve concen- ratios and d15N values for Lake Skardtjørna and Lake trations expressed in 107 valves per g dry sediment, estimated Tjørnskardet diatom diversity (standardized count size n = 100), organic 123 J Paleolimnol richness curve for Lake Skardtjørna (Fig. 4) shows percentages of N. perminuta-frustulum complex, initially low values (\5 taxa). Shortly after 1900, the and with F. pinnata becoming the dominant species. diatom richness starts to increase, a tendency that Abundances of C. placentula and N. amphibia persists until the most recent sediment samples. decrease markedly. After the dilution episode, the diatom total valve concentration continues to increase Lake Tjørnskardet to the present. For Lake Tjørnskardet, the curve shows a trend with gradually increasing values The lowermost 10 cm of the sediment core from throughout the entire sediment sequence (Fig. 4). Lake Tjørnskardet is dominated by Amphora libyca, often making up [80% (Fig. 3). From ca. 13 cm Lake Istjørna upwards, this species decreases in abundance to 5% at the top of the core. At ca. 12 cm, diatom taxa such Very few diatoms were recorded in the uppermost as Cocconeis placentula, A. inariensis, Cymbella sediments of Lake Istjørna (Fig. 3) and diatoms are (sensu lato) and Navicula costulata appear, of which completely absent below 6 cm sediment depth (prior C. placentula becomes the dominant species ([40%) to *1900). The most abundant taxa encountered between 9 and 7 cm. From ca. 7 cm (1920s) upwards, include the genera Navicula, Cymbella (sensu lato), diatoms such as F. pinnata, F. construens var. cf. Nitzschia, and Achnanthes. Interestingly, a few pumila, Nitzschia amphibia, N. perminuta-frustulum planktonic taxa were recorded in Lake Istjørna, complex and Achnanthes minutissima increase in reaching [10% in some samples, i.e. Meridion relative abundance. These taxa were completely circulare (4.0–3.0 cm), Fragilaria capucina (2.0– absent or present in very low frequencies below 0.5 cm), and Tabellaria flocculosa (1.5–1.0 cm). At 7 cm depth. The top ca. 5 cm (from *1950) is Lake Istjørna the total diatom valve concentration is dominated by F. pinnata (often [25%), Nitzschia lower than in the other lakes. On the other hand, the perminuta-frustulum complex (up to 20% in the top chrysophyte stomatocyst concentration is far higher. 2.5 cm), together with Fragilaria construens var. cf In most of the core, siliceous algae are absent, pumila, Cymbella (sensu lato), Nitzschia dissipata- possibly due to strong dissolution. Diatom and alpina-liebtruhii complex, and A. minutissima, all in chrysophyte stomatocyst concentrations show syn- abundances \10%. chronous patterns. For both indicator groups, con- Diatom total valve concentrations increased dra- centrations increase abruptly at 5 cm sediment depth matically in Lake Tjørnskardet in the top ca. 7 cm (after 1900), with high abundances of Navicula (since the 1920s), except for a short-lived decrease at species, Nitzschia species, Meridion circulare and 5 cm. In Lake Tjørnskardet, the biggest changes Cymbella (sensu lato). The increase is followed by a occur at ca. 12, 7, and 5 cm. At 12 cm (*1850), the pronounced minimum at 2.5 cm (1980). Subse- diatom assemblage, which previously consisted more quently, the concentrations for both diatoms and or less only of A. libyca and F. pinnata, changes, chrysophyte stomatocysts increase again. The chryso- with new taxa such as A. inariensis, C. placentula, phyte stomatocysts in Lake Istjørna show overall N. costulata and others. The assemblage change higher concentrations than the diatoms, even though correlates to a slight increase in total diatom valve \100 cyst specimens were identified in each sample. concentration. At 7 cm (1920) the diatom concentra- Lake Istjørna is also affected by an episode of tion starts increasing markedly. At the same time the increased sedimentation rate, between ca. 3.5 and diatom flora changes, the A. libyca abundances 1.75 cm (*1970s), an interval within which both decrease, making up\5% at the top, whereas diatoms diatom total valves and chrysophyte stomatocyst such as Nitzschia species, A. minutissima and concentrations decrease markedly. Thereafter, the F. construens var. venter and others increase in concentrations of siliceous algae increase rapidly abundance. At 5 cm (1950), the diatom total valve again. F. capucina, Achnanthes fragilaroides, Navic- concentration decreases. This is correlated with the ula schmassmannii and other Navicula species con- up-core decline in 210Pb activity (Fig. 2), due to an tinue to occur at, or increase to, relatively high episode of increased bulk sedimentation rate. There- abundances, whereas other species like M. circulare after, the assemblage changes, with increasing and Nitzschia species decrease. T. flocculosa occurs 123 J Paleolimnol at its highest abundance just after the episode with Organic carbon, nitrogen, and C:N ratios high sedimentation rate. In the uppermost *0.5 cm (late 1990s to early twenty-first century) the total The organic carbon content in the sediments is diatom valve concentration decreases, whereas the relatively low in the cores from Lake Skardtjørna and chrysophyte stomatocyst concentration continues to Lake Tjørnskardet, varying between ca. 0.9 and 1.6% increase. This interval is correlated to another episode in Lake Skardtjørna and between ca. 0.75 and 1.4% in of increased sedimentation rate. This interval is also Lake Tjørnskardet (Fig. 4). In Lake Skardtjørna, characterized by decreasing abundance of F. capuci- marked and near-synchronous drops in both organic na and others, and increasing abundance of A. minu- carbon and total nitrogen content correspond to the tissima and small Navicula species. T. flocculosa is previously mentioned 210Pb decline at ca. 10 cm completely absent in the uppermost cm of the core. depth. After this episode, both organic carbon and nitrogen show an increase to the present. The C:N molar ratio is fairly stable throughout the Lake Lake Istjørnelva Skardtjørna core, with values around eight. In Lake Tjørnskardet, both the organic carbon and nitrogen At Lake Istjørnelva (Fig. 3) a major change occurs at content show an increasing trend throughout the *5.5 cm, i.e. after the beginning of the 1900s, when twentieth century. The C:N molar ratio shows an there is a marked increase in diatom total valve increasing trend with time before 1900, reaching concentration. Before 1900, below 5.5 cm, there are values around 15. Thereafter the ratios decrease to few specimens in the core, possibly due to dissolu- values of ca. 10 to the present. tion. The chrysophyte stomatocyst concentrations in Lake Istjørnelva seem to start increasing later, after Nitrogen isotopes the 1950s. From 5.5 to 3.0 cm, Navicula digitulus is the most common species (30–50%), whereas Pin- Measurements of the nitrogen stable isotopic ratio nularia biceps, F. pinnata (up to 15%) and other (d15N) in the sediment organic matter show that the N Pinnularia and Achnanthes species are less common biogeochemistry has changed markedly in Lake members of the assemblage. Pinnularia biceps shows Skardtjørna and Lake Tjørnskardet coincident with an increasing abundance ([20%) within this interval changes in the diatom record (Fig. 4). In Lake and becomes the most dominant species between 3.0 Tjørnskardet, the sediment d15N becomes increas- and 2.5 cm (40% relative abundance). Above 2 cm, ingly depleted, by ca. 2%, in the latter half of the all Pinnularia species occur in low abundances twentieth century. The d15N values range between 3.8 (\10%). N. perminuta-frustulum complex is recorded and 4.3% before 1950, decreasing to the lowest value more frequently. In the interval 2.0–1.0 cm, other of 2.4% at the end of the twentieth century. For Lake species such as Fragilaria tenera and Stauroneis Skardtjørna, the sediment d15N values are relatively anceps appear (*10%), whereas the uppermost cm is stable before 1900, ranging between 5.3 and 5.6%, again dominated by Navicula digitulus ([30%), thereafter the d15N become increasingly depleted, to Navicula species and F. pinnata types. values around 4%. Furthermore, the d15N profile of Lake Istjørna is, much like the other lakes, affected Skardtjørna shows two inversions, corresponding to by an episode of increased sedimentation rate some- the previously mentioned 210Pb activity declines at time in the mid 1970s to late 1980s, causing dilution 10 cm and at the top of the core. of the diatom and chrysophyte stomatocyst concentrations. After the mid 1980s, both diatom and Holocene variability chrysophyte concentrations increase dramatically, particularly in the uppermost cm (after the 1990s). Pre-industrial, natural ecosystem variability was There is an apparent taxonomic change coupled to the obtained by 122 diatom analyses and 47 analyses of dilution episode. Pinnularia almost disappears and is organic carbon and C:N molar ratios from the 5.5-m taken over by increasing relative abundances of N. long core from Lake Skardtjørna (Fig. 5). The core perminuta-frustulum complex, Navicula and Fragi- covers ca. 8,000 calibrated years (the base of the core laria spp. was dated to 8,230 cal BP). A second order 123 J Paleolimnol

Fig. 5 Recent and Holocene records of organic carbon, C:N ratios and diatom total valve concentrations for Lake Skardtjørna. CRS Age model for gravity core and Calibrated Age (cal BP) for long core are presented polynomial function was fitted to the dates (r2 = exception of near the bottom, below *4.25 m. From 0.996). The core consists of silty clay at 5.5–4.0 m 5.45 to 4.15 m, dating to ca. 8,100–5,400 cal BP, the depth, silty clay with marked laminations from 4.0 to organic carbon content reaches values of up to 3%. 0.7 m depth, and silty/clayey gyttja from 0.7 m to the Within this section there is a period with low organic top (Fig. 5). Total diatom valve concentrations, carbon content, around 7,500–6,750 cal BP (ca. 5.18– organic carbon content and C:N ratios from the long 4.8 m depth). At ca. 4.15–2.46 m depth (around core and the short core from Skardtjørna were 5,400–2,650 cal BP) C:N ratios suddenly rise, reach- compared (Fig. 4). The two cores were correlated ing values up to 15. Organic carbon content decreases by comparing the geochemistry (typically low con- to ca. 1–1.5% in the beginning of the period, and centrations of Cs, Zn and Cu, but high Na2O in this stays around this level to present. At 2.46 m part of the gravity core) as well as by comparing (*2,650 cal BP) C:N suddenly decreases to values diatom records (dramatic floristic changes found in *7. Thereafter the values increase to around 10 up to both cores). ca. 0.9 m depth (*800 cal BP), with the exception of Organic carbon content is relatively low, generally the section ca. 1.64–1.24 m (ca. 1,600–1,150 cal BP), \1.5%, throughout the whole long core, with the where both the organic carbon and C:N ratios 123 J Paleolimnol decrease slightly. At ca. 0.85–0.29 m depth (*750– environmental changes on Spitsbergen in the twentieth 250 cal BP) the C:N ratio increases markedly to century. At Lake Skardtjørna, the non-Fragilaroid taxa values [15. Within this period of high C:N ratios, concentration has increased and in the other three there is a short-term interval at ca. 0.69–0.47 m (600– lakes, the total diatom valve concentrations have 400 cal BP) where the C:N ratios decrease. The increased dramatically since the beginning of the uppermost 0.29 m (most recent ca. 250 years) show 1900s. In Lake Istjørna and Lake Istjørnelva the relatively low C:N ratios. chrysophyte cyst concentrations have also increased The diatom valves are highly dissolved in a large markedly during the last century. The most pro- section of the core. From 0.33 to 4.7 m there are no nounced changes in diatom assemblages seem to have diatoms at all. However, in the top 30 cm and the occurred in the last few decades. Diatom dissolution bottom 70 cm of the long core, diatom concentration could possibly have affected the diatom flora and values are high and the valves show little or no signs abundances. At Lake Skardtjørna, dissolution is of dissolution. The diatom flora in the lower part of apparent. However, it is unlikely that F. pinnata would the core consists mainly of F. pinnata (*98–100% in remain as the only species in the lower part of the core every sample), and few other taxa were found. because of dissolution. If originally, big and robust diatoms like Amphora spp. were present, there would likely have been fragments of these left in the Discussion and conclusions sediment. While dissolution may have affected the diatom valve concentrations in the lakes to some Paleolimnological studies from around the Arctic extent, it seems likely that the floristic changes are show that dramatic shifts in lake ecosystems have more or less unaffected. We suggest that the marked occurred since around AD 1850 (Douglas et al. 1994; increases in diatom concentrations are due primarily to Birks et al. 2004; Jones and Birks 2004; Smol et al. increased primary production. 2005). These recent shifts, including changes in The taxonomic richness obtained from rarefaction species assemblages and ecological reorganizations analysis shows that the diversity of diatom assem- in siliceous algae and invertebrates, are thought to be blages in Lake Skardtjørna increased synchronously related primarily to recent climate warming (Douglas with the non-Fragilaria concentrations towards the et al. 1994, Overpeck et al. 1997, Smol et al. 2005). top of the core. In Lake Tjørnskardet the species Amounts of atmospheric nitrogen deposition are diversity has increased continuously during the considered too low to be of concern in the Arctic deposition of the sequence. (AMAP 1997) and Birks et al. (2004) concluded that there is no evidence in paleolimnological records Possible causes for the environmental changes from Spitsbergen for impacts resulting from atmospheric nitrogen deposition. However, ice cores from Climate warming is a possible cause for the diatom Lomonosovfonna (Spitsbergen) and Austfonna responses recorded in the sediments of Lake Skardt- - (Nordaustlandet) show dramatic increases of NO3 jørna and Lake Tjørnskardet, and both geological concentrations (Fig. 6) in the late twentieth century observations and meteorological records confirm (Simo˜es and Zagorodnov 2001; Kekonen et al. increasing mean annual temperatures on Svalbard 2002; Isaksson et al. 2003). In addition, evidence after the Little Ice Age (Werner 1993; Hanssen-Bauer from terrestrial ecosystems on Svalbard (Gordon and Førland 1998; Isaksen et al. 2000; Ziaja 2001). et al. 2001) show, contrary to the conclusions of the However, the question is whether the climate has AMAP report (1997), that changes caused by changed enough in the twentieth century to be the enhanced nitrogen deposition may already be occur- only explanation for the observed biota changes in ring in some Arctic ecosystems. Furthermore, studies the two lakes. of mid-latitude alpine lakes show considerable The Late Holocene Little Ice Age (LIA) on responses in lake ecosystems to recent nitrogen Svalbard is characterized by advancing glaciers enrichment (Wolfe et al. 2001, 2003). (Werner 1993; Svendsen and Mangerud 1997; Snyder The sediment cores from the four lakes in this study et al. 2000, Mangerud and Landvik 2007). In the long reveal stratigraphic variations that show dramatic core from Lake Skardtjørna (Fig. 5) the C:N molar 123 J Paleolimnol

Fig. 6 Relevant ice and paleoclimate data: Mean annual temperature data from Longyearbyen, 1912– 1997; Ground surface temperature reconstruction made from temperature measurements of a borehole in permafrost on Janssonhaugen, Svalbard; Nitrate concentration values from Austfonna and Lomonosovfonna ice cores, Svalbard

ratios suddenly increase, reaching values[15 around 1994; Meyers and Lallier-Verge`s 1999), suggesting 750 cal BP (*thirteenth century AD). Algae com- that the C:N record between ca. 700 and ca. 250 cal monly have a C:N molar ratio between 4 and 10, BP reﬂects a marked decrease of algae production while terrestrial matter has a C:N ratio [20 (Meyers in the lake, probably due to climate deterioration. 123 J Paleolimnol

Within this period there is a short interval at ca. 600– (1970) suggested that the greatest development of 400 cal BP, where the C:N ratios decrease. This two- vegetation occurred ca. 5,000 BP. The record of step change correlates well with other records on thermophilous marine molluscs in fjords in Spitsber- Spitsbergen. A study on sediment cores from Linne´- gen implies warmer conditions in the Early Holocene vatnet, northeast of Lake Skardtjørna, indicates a from around 9,500 year BP (Salvigsen et al. 1992). two-stage LIA showing a two-phase glacier growth, In this study of Lake Skardtjørna, the highest inferred from total organic carbon measurements, a organic carbon (up to 3%) content throughout the first glacial advance in the thirteenth or fourteenth whole long core, occurs around 8,100–5,400 cal BP, century AD and a second, longer-lasting glacial implying high production within the lake during this advance after AD 1500, culminating in the nineteenth period (Fig. 6). C:N ratios are \10, suggesting that century (Svendsen and Mangerud 1997). the major source of the organic carbon in this part of In the first half of the twentieth century there was a the core is lacustrine, i.e. from algae production. It is general glacial retreat in Svalbard, in response to also here in the lowermost part of the long core, warming climate (Werner 1993, Ziaja 2001). A corresponding to ca. 8,000–6,600 cal BP, that dia- ground surface temperature reconstruction made from toms occur in high concentrations. The flora during temperature measurements in a permafrost borehole this period with warmer conditions consists almost in Janssonhaugen, Svalbard (Fig. 6), shows clear exclusively of F. pinnata, although other taxa occur warming of surface temperatures over the last in very low numbers. 60–80 years (Isaksen et al. 2000). Meteorological In addition to climate warming, increased precip- observations in Longyearbyen from 1912 to 1996 itation might also have contributed to the changes (Førland et al. 1997) show that the annual mean seen in the lakes in the twentieth century. In all four temperature increased from 1912 to the late 1930s lakes there are marked decreases in diatom concen- and from the 1960s to 1996 (Fig. 5). Between these trations, as well as in the chrysophyte cyst concen- warming periods there were decreasing temperature tration at Istjørna and Istjørnelva, corresponding to trends. Spring is the only season that shows a the declines seen in the 210Pb activity graphs (Fig. 2). warming trend during the period 1912–1996 as a The decreases are probably caused by episodes of whole (Hanssen-Bauer and Førland 1998). Further, increased sedimentation rates during which enhanced the present temperature is about the same as in the input of inorganic sediment diluted the concentration 1920s, but slightly lower than during the 1930s and of siliceous algae as well as the 210Pb activity, as 1950s (Hanssen-Bauer and Førland 1998). described in Wolfe et al. (2004). The dilution Climate warming could affect primary production episodes correlate with major reorganizations in in the Svalbard lakes by prolonging the growing diatom assemblages in all four lakes. Changes in period induced by a longer ice-free season. Consid- the catchments such as increased inflow of sediment ering the short ice-free period during the summer, it is to the lakes are likely to change both immediate and likely that even a small change in temperature leading more long-lasting growing and habitat conditions for to decreased lake ice-cover would affect algae diatoms, i.e. changing light conditions due to growth. We suggest that climate warming has indeed increased suspended sediment in the water column affected the production of siliceous algae in the past. as well as changing nutrient availability, pH, etc. During the early and mid-Holocene, the summer The precipitation data from Svalbard airport temperatures on Spitsbergen were significantly higher are not without uncertainties, but the mean annual than today (Svendsen and Mangerud 1992; Svendsen precipitation seems to have increased by ca. 30% et al. 1996). Around 10,000 BP, a rapid climatic during the period 1912–1996 (Førland et al. 1997; warming occurred, causing some cirque and valley Hanssen-Bauer and Førland 1998). There also seems glaciers to disappear. A previous study of Lake to have been a period with increased precipitation Skardtjørna sediments suggested a mean July tem- between 1965 and 1975 (Førland et al. 1997; perature at least 1.5°C higher than present (Birks Hanssen-Bauer and Førland 1998). 1991) around 8,000–4,000 BP (uncalibrated 14C Increasing diatom concentrations in Spitsbergen in years), based on the occurrence of macrofossils of the early twentieth century could possibly be ascribed Cassiope hypnoides and Salix cf. glauca. Hyva¨rinen to climate warming following the end of the Little Ice 123 J Paleolimnol

Table 3 d15N values on modern and Holocene leaves of Salix polaris, sampled in Adventdalen and in Lake Skardtjørna Sample type Location Depth Isotope ratio Age Summary d15N(% air)

Salix polaris leaf Modern Adventdalen Bag 1, sample 1 -4.87 Modern Salix polaris leaf Modern Adventdalen Bag 2, sample 2 -3.16 Modern Salix polaris leaf Modern Adventdalen Bag 1, sample 3 -5.02 Modern Mean -4.18 Salix polaris leaf Modern Adventdalen Bag 2, sample 4 -4.7 Modern SD 0.64 Salix polaris leaf Modern Skardtjørna AA-1 -3.7 Modern Salix polaris leaf Modern Skardtjørna AA-2 -4.14 Modern Salix polaris leaf Modern Skardtjørna BB-1 -4.00 Modern Salix polaris leaf Modern Skardtjørna BB-2 -3.89 Modern Salix polaris leaf Holocene (H. Birks) 210–215 cm -2.52 *4,500 Salix polaris leaf Holocene (H. Birks) 265–270 cm -0.93 *5,500 Salix polaris leaf Holocene (H. Birks) 315–320 cm 0.17 *7,000 Mean -1.33 Salix polaris leaf Holocene Skardtjørna 4,841–4,863 cm -1.01 6,770–6,820 SD 0.95 Salix polaris leaf Holocene Skardtjørna 4,885–4,907 cm -1.92 6,870–6,910 Salix polaris leaf Holocene Skardtjørna 4,977 cm -1.78 7,060 Dd15N -2.85

Age. The warming, especially spring temperatures, possibly years (Lehmann et al. 2002). In this study, the may have caused prolonged ice-free periods and longer d15N trends are expressed on a timescale of several growing seasons. However, the temperature trends decades. Therefore, it is reasonable to assume that the suggested by the meteorological data from Longyear- d15N trends we show here primarily reflect isotopic byen, do not convincingly support temperature variations in the overlying water column, as suggested increase as the only explanation for the dramatic biota elsewhere (Teranes and Bernasconi 2000; Wolfe et al. changes in the lakes during the last few decades. 2003). We suggest that the low values of the modern Furthermore, the striking taxonomic richness in the Salix polaris leaves, as well as the generally decreasing past few decades, compared to the early Holocene d15N trends in sediments from Skardtjørna and sparse flora, also indicates another explanation. Tjørnskardet, primarily reflect a supply of isotopically In both Lake Skardtjørna and Lake Tjørnskardet, the depleted N from long-distance atmospheric sources. nitrogen biogeochemistry changed simultaneously Anthropogenic N, from sources other than human and with the diatom assemblages, i.e. the nitrogen isotopic animal waste, is in general depleted compared to composition (d15N) of the sediment organic matter natural sources (Talbot 2001). For example, measure- became depleted by *2% in the second half of the ments of nitrogen emissions from coal combustion twentieth century. Furthermore, we measured d15Non show d15N values ranging from -10 to 0% (Heaton modern and early/mid Holocene leaves of Salix polaris 1990). Thus, an increased contribution of nitrogen (Table 3) from Skardtjørna and Adventdalen (SE of from anthropogenic sources is likely to cause depleted Longyearbyen). The d15N values of the modern leaves lake sediment d15N values. However, a second possible (mean -4.18) were much lower than those of the explanation for the depleted d15N values is physiolog- Holocene leaves (mean -1.33). ical fractionation during nitrogen assimilation. The Diagenetic processes could affect the d15N trends 15N/14N ratio has been used to infer past primary seen in the sediments. Diagenetic loss of nitrogen productivity changes in paleoceanographic studies would show falling N content, rising C/N values and (Altabet and François 1994). Phytoplankton discrim- 15 15 14 - enriched d N values with depth in the sediments inate against N, favouring N during NO3 assim- (Ostrom et al. 1998; Talbot 2001). However, it has ilation. In a closed system, Raleigh fractionation been shown that early diagenesis typically reaches dictates that isotopic discrimination decreases as completion on a time span of weeks to months or the DIN (dissolved inorganic nitrogen) pool is 123 J Paleolimnol

- Table 4 Spitsbergen precipitation events with associated NO3 concentrations - - Date Event NO3 NO3 Location Wind direction Wind Temperature type (lgl-1) (lM) speed (°C) (m s-1)

Apr 21, 2001 Fresh snow 124 2.0 Ny A˚ lesund N (10°) 2.7 -11.4 Apr 5, 2002 Fresh snow 248 4.0 Ny A˚ lesund WSW (260°) 2.9 -2.6 July 26, 2004 Rain 212 3.4 Kapp Linne´ SSE to WSW 3.0 10.5 (160–240°) July 29, 2004 Rain 1,271 20.5 Kapp Linne´ SSW (180– 6.4 5.2 190°) Aug 4, 2004 Rain 163 2.6 Kapp Linne´ SSW to WSW 2.0 6.3 (190–250°) Aug 8, 2004 Rain 1,561 25.2 Kapp Linne´ SSW (215°) 2.6 4.4 Feb 4, 2005 Fresh snow 174 2.8 Storfjorden WSW (240°) 1.7 -9.1 Feb 6, 2005 Fresh snow 61 1.0 Storfjorden ESE (120°) 2.8 -16.3 July 21–22, 2005 Rain 153 2.5 Kapp Linne´ NNE to WSW 3.8 3.7 (20–250°) July 28, 2005 Rain 63 1.0 Kapp Linne´ SE to WSW 2.5 4.1 (135–250°) Climate data for the two earliest dates are from Svalbard airport (29 m asl); other data are from the Gruvfjellet station (464 m asl) progressively utilized, until the fractionation has no transport of nitrogen from western parts of Europe, for observable impact under N-limited conditions (Goer- these speciﬁc episodes. Contamination from coal icke et al. 1994). Thus, in a marine environment where mining on Svalbard would be another possible source N is a limiting nutrient, the sinking organic matter for enhanced nitrogen deposition. However, data from 15 - - becomes increasingly enriched in N as the NO3 is 10 precipitation events with associated NO3 concen- progressively depleted from surface waters (Goericke trations, measured in Western Spitsbergen in 2001– et al. 1994). With the increased deposition of atmo- 2005, combined with climate data from Svalbard spheric N in the lakes in Spitsbergen, algae competition Lufthavn and Gruvfjellet station, shows that the events - for nitrogen would reduce, possibly causing sustained with the highest NO3 concentrations are associated fractionation against d15N and depletion of d15Nin with SSW wind directions (Table 4). This further sediment organic matter. However, these fractionation supports our hypothesis of a western European source effects have not been inferred as successfully in for the transport of anthropogenic nitrogen to Lake lacustrine environments (Teranes and Bernasconi Skardtjørna and Lake Tjørnskardet. 2000). In fact, many recent freshwater studies support To identify the anthropogenic contributions to N the use of sediment d15N as an indicator of source N changes in the lakes, a simpliﬁed, two-component isotopic composition (Teranes and Bernasconi 2000; isotope mixing model was developed (Fig. 7) (Phillips Wolfe et al. 2003). and Gregg 2001; Wolfe et al. 2003), with the equations: 15 15 15 Anthropogenic nitrogen in Lake Skardtjørna and d Nlake = (fnat d Nnat) ? (fanthro d Nanthro) and 15 Lake Tjørnskardet, could be from both local and long- fnat ? fanthro = 1.00, so that fanthro = (d Nlake - 15 15 15 distance sources. Trajectory climatology developed for d Nnat)/(d Nanthro - d Nnat), where the isotopic 15 Svalbard (Eneroth et al. 2003), presents eight major signature of the organic matter in the lake (d Nlake) 15 atmospheric transport patterns to Ny-A˚ lesund, includ- combines input from both natural (d Nnat) and 15 ing transport of air masses from the Eurasian continent. anthropogenic sources (d Nanthro), each represented A case study of the back-trajectories for episodes of by their respective fractions (fnat and fanthro) contrib- 15 high concentrations of nitrate in a stream at Ny- uted to d Nlake. Analyses of nitrate in snow from A˚ lesund in 1997 and 1999, suggests an atmospheric Ny-A˚ lesund, Svalbard, in the years 2001–2003, show

123 J Paleolimnol

Fig. 7 Modeled fraction (%) of anthropogenic nitrogen using data from Ny A˚ lesund (Heaton et al. 2004) and the stable isotope measurements

low d15N values, ranging from -18 to -7% (versus turnover as beta-diversity, scaled in standard devia- air N2; Heaton et al. 2004), with a mean value of tion (SD) units and obtained from detrended canon- -10.4 ± 1.2. This value was used to constrain ical correspondence analysis (DCCA), and compared 15 d Nanthr (Heaton et al. 2004). The most enriched results with values for four other lakes on Spitsbergen values near the bottom of the cores were used for (Smol et al. 2005). Whereas the lakes in this study 15 d Nnat. Results show that 10–15% of N subsidies to show a species turnover ranging from 1.84 to 2.37 SD these lakes in the twentieth century appear to be units, the additional lakes on Spitsbergen reveal anthropogenic in origin. The changes in nitrogen considerably lower species turnover from 0.89 to 1.40 availability in the latter half of the twentieth century SD units (Table 5). Following the study by Smol are compatible with the recent increases in algal et al. (2005), we also compared our results with the production. Increased nitrogen availability also cor- results from other lake populations analyzed in the responds to the diversification in diatom assemblages same way. The beta-diversities of our lakes were in the past few decades. For example, the domi- normalized to the average values from 14 diatom nance of Fragilaria species in Lake Skardtjørna records from temperate, relatively unaffected refer- in the lower part of the short core (Fig. 3) and in ence lakes in Scotland, Canada and Ireland (Smol the Holocene diatom record (Fig. 5) may reflect a et al. 2005 and references therein), as well as to the possible N-limited environment in the early history of average values from 42 circumpolar lakes (Smol et al. Lake Skardtjørna, whereas the decrease of Fragilaria 2005) and 14 strongly acidified lakes in Norway, spp. and the sudden diversification of diatom species Sweden and The United Kingdom (Smol et al. 2005). may reflect elevated nutrient concentrations in recent Diatom species compositional turnover since 1850 is decades. about twice as high in our lakes than in undisturbed temperate lakes, and it is comparable in magnitude to Comparison with other lakes changes observed in acidified lakes. Moreover, the lakes in our study show much greater beta diversities The timing of the ecological shifts recorded in our than the lakes in the study by Jones and Birks (2004). lakes is comparable to observations from other lakes This stresses the fact that ecological responses may on Spitsbergen (Jones and Birks 2004). For the vary markedly between lakes situated within a four lakes in this study, we estimated total species relatively small geographic area. 123 J Paleolimnol

Table 5 Lake locations and beta-diversity (species turnover) estimates since *AD 1850 for the eight available Svalbard diatom records Site name Latitude Longitude Beta- Beta-diversity/ Beta-diversity/ Beta-diversity/ Beta-diversity/ diversity temperate reference northern reference arctic reference acidiﬁed reference (SD units) (n = 14) (n = 42) (n = 27) (n = 14)

Istjørna 78°020N15°080E 2.37 2.42 2.14 1.96 1.20 Skardtjørna 77°570N13°490E 2.12 2.16 1.91 1.75 1.07 Istjørnelva 78°000N15°140E 1.90 1.94 1.71 1.57 0.96 Tjørnskardet 77°580N13°520E 1.84 1.88 1.66 1.52 0.93 Birgervatnet 79°480N11°370E 1.40 1.43 1.26 1.16 0.71 Ytertjørna 78°130N12°560E 1.15 1.17 1.04 0.95 0.58 Arresjøen 79°400N10°510E 1.07 1.09 0.96 0.88 0.54 ‘‘Scurvy 79°440N12°180E 0.89 0.91 0.80 0.74 0.45 Pond’’ Beta-diversity is also shown normalized to the average reference values from additional lake populations analyzed in the exact same manner

There is, however, another factor that might have AMAP (1998) Arctic monitoring and assessment program: contributed to differences in beta diversity. The lakes Arctic pollution issues: a state of the Arctic environment report. Oslo in our study were cored in 2001, while Ytertjørna was Appleby PG, Oldfield F (1978) The calculation of lead-210 cored in 1995 and Arresjøen, Birgervatnet and dates assuming a constant rate of supply of unsupported Scurvy Pond were sampled in 1993. This means 210Pb to the sediment. Catena 5:1–8. doi:10.1016/S0341- there is a time difference of 6–8 years between the 8162(78)80002-2 Battarbee RW, Kneen MJ (1982) The use of electronically studies, a time period during which dramatic counted microspheres in absolute diatom analysis. Limnol increases in diversity and complexity of diatom Oceanogr 27:184–188 communities continued to occur. Berger A, Loutre MF (1991) Insolation values for the climate We conclude that the recent increases of algal of the last 10000000 years. Quat Sci Rev 10:297–317. doi:10.1016/0277-3791(91)90033-Q production seen in Lake Skardtjørna and Lake Bergstro¨m A-K, Jansson M (2006) Atmospheric nitrogen Tjørnskardet most likely resulted from the combined deposition has caused nitrogen enrichment and eutrophi- effects of warming climate, i.e. longer summer cation of lakes in the northern hemisphere. Glob Change growing season, and increased nutrient availability Biol 12:635–643. doi:10.1111/j.1365-2486.2006.01129.x Birks HH (1991) Holocene vegetational history and climate primarily due to an enhanced contribution of nitrogen change in west Spitsbergen–plant macrofossils from from atmospheric deposition. Skardtjørna, an Arctic lake. Holocene 1:209–215. doi: 10.1177/095968369100100303 Acknowledgments We thank the Swedish Research Council Birks HJB, Line JM (1992) The use of rarefaction analysis for and Ymer-80 for supporting this study, Jørn Dybdahl for estimating palynological richness from the Quaternary logistics, Jack Cornett and Elis Holm for sediment chronology, pollen-analytical data. Holocene 2:1–10 and Eric Steig for isotopic measurements. We thank Professor Birks HJB, Jones VJ, Rose NL (2004) Recent environmental Hilary H. Birks for providing plant macrofossil samples. change and atmospheric contamination on Svalbard as Professor Barbara Wohlfarth is gratefully acknowledged for recorded in lake sediments–synthesis and general con- advice and taking part in fieldwork. clusions. J Paleolimnol 31:531–546. doi:10.1023/B:JOPL. 0000022550.81129.1a Brattbakk I (1986) Vegetasjonsregioner—Svalbard og Jan References Mayen. Nasjonalatlas for Norge, Kartblad 4. 1. 3 Comiso JC (2003) Warming trends in the Arctic from Altabet MA, François R (1994) Sedimentary nitrogen isotopic clear sky satellite observations. J Clim 16:3498–3510. ratio as a recorder for surface ocean nitrate utilization. doi:10.1175/1520-0442(2003)016\3498:WTITAF[2.0. Global Biogeochem Cycles 8:103–116. doi:10.1029/ CO;2 93GB03396 Douglas MSV, Smol JP, Blake W Jr (1994) Marked post– AMAP (1997) Arctic monitoring and assessment programme: eighteenth century environmental change in High–Arctic Arctic pollution issues: a state of the Arctic environment ecosystems. Science 266:416–419. doi:10.1126/science. report. Oslo 266.5184.416

123 J Paleolimnol

Eneroth K, Kjellstro¨m E, Holmeń K (2003) A trajectory cli- Jones VJ, Birks HJB (2004) Lake-sediment records of recent matology for Svalbard; investigating how atmospheric environmental change on Svalbard: results of diatom flow patterns influence observed tracer concentrations. analysis. J Paleolimnol 31:445–466. doi:10.1023/B:JOPL. Phys Chem Earth 28:1191–1203 0000022544.35526.11 Foged N (1964) Freshwater diatoms from Spitsbergen. Tromsø Kaushal S, Binford MW (1999) Relationship between C:N Museums Skrifter, 205 pp ratios of lake sediments, organic matter sources, and Førland E, Hanssen-Bauer I, Nordli Ø (1997) Climate statistics historical deforestation in Lake Pleasant, Massachusetts, and long-term series of temperature and precipitation at USA. J Paleolimnol 22:439–442. doi:10.1023/A:1008027 Svalbard and Jan Mayen. DNMI Report 21/97 KLIMA, 028029 Norwegian Meteorological Institute, Oslo Kekonen T, Moore J, Mulvaney R, Isaksson E, Pohjola V, van Gajewski K, Hamilton PB, Mcneely R (1997) A high resolution de Wal RSW (2002) A 800 year record of nitrate from the proxy-climate record from an arctic lake with annually Lomonosovfonna ice core, Svalbard. Ann Glaciol 35:261– laminated sediments on Devon Island, Nunavut, Canada. J 265. doi:10.3189/172756402781817121 Paleolimnol 17:215–225. doi:10.1023/A:1007984617675 Koerner RM, Fisher DA (2002) Ice-core evidence for wide- Glew JR (1989) A new trigger mechanism for sediment sam- spread Arctic glacier retreat in the Last Interglacial and plers. J Paleolimnol 2:241–243. doi:10.1007/BF00195474 the early Holocene. Ann Glaciol 35:19–24. doi:10.3189/ Glew JR, Smol JP, Last WM (2001) Sediment core collection 172756402781817338 and extrusion. In: Last WM, Smol JP (eds) Tracking Krammer K, Lange-Bertalot H (2001) Bacillariophyceae. In: environmental change using lake sediments volume 1 Ettl H, Gerloff H, Heynig H, Mollenhauer D (eds) Su¨s- Basin analysis, coring and chronological techniques. swasserflora Von Mitteleuropa 2/1–4. Gustav Fischer Kluwer Academic Publishers, Dordrecht, pp 73–105 Verlag, Stuttgart Goericke R, Montoya JP, Fry B (1994) Physiology of isotopic Landvik JY, Mangerud J, Salvigsen O (1987) The late fractionation in algae and cyanobacteria. In: Lajtha K, Weichselian and Holocene shoreline displacement on the Michener RH (eds) Stable isotopes in ecology, environ- west-central coast of Svalbard. Polar Res 5:29–44. mental science. Blackwell, Boston, pp 187–221 doi:10.1111/j.1751-8369.1987.tb00353.x Gordon C, Wynn JM, Woodin SJ (2001) Impacts of increased Lehmann MF, Bernasconi SM, Barbieri A, McKenzie JA (2002) nitrogen supply on high Arctic heath: the importance of Preservation of organic matter and alteration of its carbon bryophytes and phosphorous availability. New Phytol and nitrogen isotope composition during simulated and in 149:461–471. doi:10.1046/j.1469-8137.2001.00053.x situ early sedimentary diagenesis. Geochim Cosmochim Goto-Azuma K, Koerner RM (2001) Ice core studies of Acta 66:3573–3584. doi:10.1016/S0016-7037(02)00968-7 anthropogenic sulphate and nitrate trends in the Arctic. J Major H, Nagy J (1972) Geology of the Adventdalen map area. Geophys Res 106:4959–4969. doi:10.1029/2000JD900635 Norsk Polarinstitutt Skrifter Nr 138 Hanssen-Bauer I, Førland EJ (1998) Long-term trends in pre- Mangerud J, Landvik JY (2007) Younger Dryas cirque glaciers in cipitation and temperature in the Norwegian Arctic: can western Spitsbergen: smaller than during the Little Ice Age. they be explained by changes in atmospheric circulation Boreas 36:278–285. doi:10.1080/03009480601134827 patterns? Clim Res 10:143–153. doi:10.3354/cr010143 Mayewski PA, Lyons WB, Spencer MJ, Twickler MS, Buck 15 14 Heaton THE (1990) N/ N ratios of NOx from vehicle engines SI, Whitlow S (1990) An ice-core record of atmospheric and coal-fired power stations. Tellus 42B:304–307 response to anthropogenic sulfate and nitrate. Nature Heaton THE, Wynn P, Tye AM (2004) Low 15N/14N ratios for 346:554–556. doi:10.1038/346554a0 nitrate in snow in the High Arctic (79°). Atmos Environ Meyers PA (1994) Preservation of elemental and isotopic 38:5611–5621. doi:10.1016/j.atmosenv.2004.06.028 source identification of sedimentary organic matter. Chem Hjelle A, Lauritzen Ø, Salvigsen O, Winsnes TS (1986) Geo- Geol 114:289–302. doi:10.1016/0009-2541(94)90059-0 logical map, Svalbard 1: 100.000, sheet B10G Van Mi- Meyers PA, Lallier-Verge`s E (1999) Lacustrine organic matter jenfjorden. Norsk Polarinstitutt Temakart No 2 records of Late Quaternary paleoclimates. J Paleolimnol Hobbie JE, Peterson BJ, Bettez N, Deegan L, O’Brien WJ, 21:345–372. doi:10.1023/A:1008073732192 Kling GW, Kipphut GW, Bowden WB, Hershey AE Ostrom NE, Long DT, Bell EM, Beals T (1998) The origin and (1999) Impact of global change on the biogeochemistry cycling of particulate and sedimentary organic matter and and ecology of an arctic freshwater system. Polar Res nitrate in Lake Superior. Chem Geol 152:13–28. doi: 18:207–214. doi:10.1111/j.1751-8369.1999.tb00295.x 10.1016/S0009-2541(98)00093-X Hyva¨rinen H (1970) Flandrian pollen diagrams from Svalbard. Overpeck J, Hughen K, Hardy D, Bradley R, Casse R, Douglas Geogr Ann 52A:213–222. doi:10.2307/520815 M, Finney B, Gajewski K, Jacoby G, Jennings A, Isaksen K, Vonder Mu¨hll D, Gubler H, Kohl T (2000) Ground Lamoreux S, Lasca A, MacDonald G, Moore J, Retelle M, surface temperature reconstruction based on data from a deep Smith S, Wolfe A, Zielinski G (1997) Arctic environ- borehole in permafrost at Janssonhaugen, Svalbard. Ann mental change of the last four centuries. Science Glaciol 31:287–294. doi:10.3189/172756400781820291 278:1251–1256. doi:10.1126/science.278.5341.1251 Isaksson E, Hermanson M, Hicks S, Igarashi M, Kamiyama K, Phillips DL, Gregg JW (2001) Uncertainty in source parti- Moore J, Motoyama H, Muir D, Pohjola V, VaikmaëR, tioning using stable isotopes. Oecologia 127:171–179. van de Wal RSW, Watanabe O (2003) Ice cores from doi:10.1007/s004420000578 Svalbard–useful archives of past climate and pollution Salvigsen O, Forman SL, Miller GH (1992) Thermophilous history. Phys Chem Earth 28:1217–1228 molluscs on Svalbard during the Holocene and their

123 J Paleolimnol

paleoclimatic implications. Polar Res 11:1–10. doi: Teranes JL, Bernasconi SM (2000) The record of nitrate uti- 10.1111/j.1751-8369.1992.tb00407.x lization and productivity limitation provided by d15 N Serreze MC, Walsh JE, Chapin III FS, Osterkamp T, Dy- values in lake organic matter—A study of sediment trap urgerov M, Romanovsky V, Oechel WC, Morison J, and core sediments from Baldeggersee, Switzerland. Zhang T, Barry G (2000) Observational evidence of Limnol Oceanogr 45:801–813 recent change in the northern high-latitude environment. ter Braak CJF (1986) Canonical correspondence analysis: a Clim Change 46:159–207 new eigenvector method for multivariate direct gradient Sickman JO, Melack JM, Clow DW (2003) Evidence for analysis. Ecology 67:1167–1179 nutrient enrichment of high-elevation lakes in the Sierra ter Braak CJF, Smilauer P (2002) CANOCO reference manual Nevada, California. Limnol Oceanogr 48:1885–1892 and CanoDraw for Windows user’s guide: software for Simo˜es JC, Zagorodnov VS (2001) The record of anthropo- Canonical Community Ordination (Version 4.5). Micro- genic pollution in snow and ice in Svalbard, Norway. computer Power, Ithaca, New York Atmos Environ 35:403–413. doi:10.1016/S1352-2310(00) ter Braak CJF, Verdonschot PFM (1995) Canonical corre- 00122-9 spondence analysis and relatedmultivariate methods in Smol JP, Wolfe AP, Birks HJB, Douglas MSV, Jones V, aquatic ecology. Aquat Sci 57:255–289 Korhola A, Pienitz R, Ru¨hland K, Sorvari S, Antoniades van de Vijver B, Ledeganck P, Potters G, Beyens L (1999) D, Brooks SJ, Fallu M, Hughes M, Keatley BE, Laing TE, Diatom communities from alkaline environments of the Michelutti N, Nazarova L, Nyman M, Paterson AM, Brøgger peninsula, north–west Spitsbergen. Nova Perren B, Quinlan R, Rautio M, Saulnier-Talbot E, Sii- Hedwigia 68:93–115 tonen S, Solovieva N, Weckstro¨m J (2005) Climate-driven Werner A (1993) Holocene moraine chronology, Spitsbergen; regime shifts in the biological communities of arctic Svalbard: lichenometric evidence for multiple Neoglacial lakes. Proc Natl Acad Sci USA 102:4397–4402. doi: advances in the Arctic. Holocene 3:128–137. doi:10.1177/ 10.1073/pnas.0500245102 095968369300300204 Snyder JA, Werner A, Miller GH (2000) Holocene cirque Wolfe AP (1997) On diatom concentrations in lake sediments: glacier activity in western Spitsbergen, Svalbard: sedi- Results of an inter-laboratory comparison and other ment records from proglacial Linne´vatnet. Holocene experiments performed on a uniform sample. J Paleo- 10:555–563. doi:10.1191/095968300667351697 limnol 18:261–268. doi:10.1023/A:1007937300347 Sorvari S, Korhola A, Thompson R (2002) Lake diatom Wolfe AP, Perren BB (2001) Chrysophyte microfossils record responses to recent Arctic warming in Finnish Lapland. marked responses to recent environmental changes in Glob Change Biol 8:153–164. doi:10.1046/j.1365-2486. high- and mid-arctic lakes. Can J Bot 79:747–752. doi: 2002.00463.x 10.1139/cjb-79-6-747 Stuiver M, Reimer PJ, Bard E, Beck JW, Burr GS, Hughen KA, Wolfe AP, Baron JS, Cornett RJ (2001) Anthropogenic nitro- Kromer B, McCormac G, van der Plicht J, Spurk M gen deposition induces rapid ecological changes in alpine (1998) IntCal98 radiocarbon age calibration, 24,000–0 cal lakes of the Colorado Front Range (USA). J Paleolimnol BP. Radiocarbon 40:1041–1083 25:1–7. doi:10.1023/A:1008129509322 Svendsen JI, Mangerud J (1992) Paleoclimatic inferences from Wolfe AP, Van Gorp AC, Baron JS (2003) Recent ecological glacial ﬂuctuations on Svalbard during the last and biogeochemical changes in alpine lakes of Rocky 20000 years. Clim Dyn 6:213–220 Mountain National Park (Colorado, USA): a response to Svendsen JI, Mangerud J (1997) Holocene glacial and climatic anthropogenic nitrogen deposition. Geobiology 1:153– variations on Spitsbergen, Svalbard. Holocene 7:45–57. 168. doi:10.1046/j.1472-4669.2003.00012.x doi:10.1177/095968369700700105 Wolfe AP, Miller GF, lsen CA, Forman SL, Doran PT, Svendsen JI, Elverhoi A, Mangerud J (1996) The retreat of the Holmgren SU (2004) Geochronology of high latitude lake Barents Sea Ice Sheet on the western Svalbard margin. sediments. In: Pienitz R, Douglas MSV, Smol JP (eds) Boreas 25:244–256 Long-term environmental change in the Arctic and Ant- Talbot MR (2001) Nitrogen isotopes in paleolimnology. In: arctic lakes. Kluwer, Dordrecht, pp 1–31 Last WM, Smol JP (eds) Tracking environmental change Ziaja W (2001) Glacial recession in Sorkappland and central using Lake Sediments volume 2: physical and geochem- Nordenskio¨ldland, Spitsbergen, Svalbard, during the ical methods. Kluwer Academic Publishers, Dordrecht, pp twentieth century. Arct Antarct Alp Res 33:36–41. doi: 401–439 10.2307/1552275

123