Early-20Th Century Environmental Changes Inferred Using Subfossil Diatoms from a Small Pond on Melville Island, N.W.T., Canadian High Arctic

Home , Ellef Ringnes Island, Meighen Island, Penny Ice Cap

Hydrobiologia (2006) 553:15–26 Springer 2006 DOI 10.1007/s10750-005-1737-5

Primary Research Paper Early-20th century environmental changes inferred using subfossil diatoms from a small pond on Melville Island, N.W.T., Canadian high Arctic

Bronwyn E. Keatley1,*, Marianne S.V. Douglas2 & John P. Smol1 1Department of Biology, Paleoecological Environmental Assessment and Research Laboratory, QueenÕs University, 116 Barrie St., K7L 3N6, Kingston, Ontario, Canada 2Department of Geology, Paleoenvironmental Assessment Laboratory, University of Toronto, 22 Russell St., M5S 3B1, Toronto, Ontario, Canada (*Author for correspondence: E-mail: [email protected])

Received 22 September 2004; in revised form 18 January 2005; accepted 26 January 2005

Key words: diatoms, environmental change, high arctic, paleolimnology, Melville Island

Abstract Diatom-based paleolimnological studies are being increasingly used to track long-term environmental change in arctic regions. Little is known, however, about the direction and nature of such environmental changes in the western Canadian high Arctic. In this study, shifts in diatom assemblages preserved in a 210Pb-dated sediment core collected from a small pond on Melville Island, N.W.T., were interpreted to record marked environmental changes that had taken place since the early 20th century. For most of the history of the pond recorded in this core, the diatom assemblage remained relatively stable and was dominated by Fragilaria capucina. A major shift in species composition began in the early-20th century, with a sharp decline in F. capucina and a concurrent increase in Achnanthes minutissima. In the last 20 years, further changes in the diatom assemblage occurred, with a notable increase in the Nitzschia perminuta complex. The assemblage shifts recorded at this site appear to be consistent with environmental changes triggered by recent climatic warming.

Introduction may provide an effective alternative method of gathering records of past environmental conditions The Arctic is well-recognized to be especially when traditional monitoring data are not available sensitive to environmental change (Serreze et al., (Smol, 2002). 2000; Houghton et al., 2001). Due to a number of Unlike many other arctic proxy records, lakes feedback mechanisms, such as snow cover-albedo, and ponds are abundant throughout the Canadian proposed temperature increases are likely to be high Arctic, and thus offer the potential of greater maximized in high-latitude regions. Thus the regional synthesis. Diatoms (class Bacillariophy- Arctic comprises a critical reference area for ceae), siliceous unicellular algae, are particularly environmental change. In the Canadian high useful paleoenvironmental indicators because they Arctic, logistical constraints and the short duration are ubiquitous, they respond rapidly to changing and poor spatial coverage of the few meteorologi- conditions, and different species often have distinct cal records makes monitoring of this vast area optima to given environmental variables (Stoermer particularly difficult. The lack of long-term & Smol, 1999). instrumental data in many ways precludes an Observational data (Serreze et al., 2000) and accurate assessment of long-term environmental proxy records (e.g., Kaufman et al., 2004) indicate change. Paleolimnological techniques, however, that the timing and nature of environmental 16 changes are not synchronous across the Arctic. century, when the diatom assemblages underwent For example, divergences in stable oxygen isotope, substantial changes that were attributed to climatic atmospheric dust, and glaciochemical records warming (Douglas et al., 1994). Since this initial between ice core records from the Devon Ice Cap study, other high arctic paleolimnological investi- (Devon Island), Penny Ice Cap (Baffin Island), and gations have shown similar changes in diatom Greenland (Camp Century and GISP2) suggest community structure since 1850 (e.g., Doubleday increasingly regional climatic influences during the et al., 1995; Gajewski et al., 1997; Wolfe, 2000; Holocene (Paterson et al., 1977; Fisher, 1979; Perren et al., 2003; Michelutti et al., 2003a; OÕBrien et al., 1995; Zdanowicz et al., 2000; Antoniades et al., 2005). These shifts have not been Grumet et al., 2001). simultaneous, but rather appear to be at least To date, most paleolimnological records from partly related to the local limnological conditions, the Canadian high Arctic are from eastern regions, such as lake size (e.g., Doubleday et al., 1995; and have illustrated differences in the timing and Michelutti et al., 2003a) and other variables. For magnitude of environmental change (e.g., Douglas example, diatom assemblages from small water et al., 1994; Doubleday et al., 1995; Perren et al., bodies near Isachsen, Ellef Ringnes Island (Fig. 1, 2003; Michelutti et al., 2003a; Antoniades et al., site b) experienced species turnover starting in the 2005). For example, diatom assemblages from mid-19th century, while those from a larger lake at shallow ponds on Cape Herschel, east-central Alert, Ellesmere Island (Fig. 1, site c), only began Ellesmere Island (Fig. 1, site g), remained relatively to shift after the mid-to late 20th century static and were interpreted to record cool temper- (Antoniades et al., 2005). Likewise, subtle diatom atures for several millennia up until the mid-19th assemblage changes at a large and deep high arctic

Figure 1. Map showing location of sites discussed in this paper. The oval on the inset map shows location of Canadian High Arctic. Sites are as follows: (a) MV-AT, Melville Island; (b) Isachsen, Ellef Ringnes Island; (c) Alert, Ellesmere Island; (d) Tuborg Lake, Ellesmere Island; (e) Fosheim Peninsula, Ellesmere Island; (f) Devon Ice Cap, Devon Island; (g) Cape Hershel, Ellesmere Island; (h) Agassiz Ice Cap, Ellesmere Island; (i) Melville Island ice caps, Melville Island; (j) Meighen Ice Cap, Meighen Island; (k) Char Lake, Resolute Bay, Cornwallis Island; (l) Mould Bay, Prince Patrick Island; (m) Bathurst Island; (n) Victoria Island. 17 lake (Char Lake, Cornwallis Island, Fig. 1, site k) Table 1. Present-day physical and chemical characteristics of only began to occur in the late-1980s (Michelutti pond MV-AT were collected on July 24, 2002 et al., 2003a). These diatom community shifts were Parameter Value correlated to warmer temperatures documented by nearby instrumental meteorological records. Latitude 75 19¢ N Warming conditions are expected to result in a Longitude 111 25¢ W lengthening of the growing season and enhanced Maximum depth 0.40 m autochthonous production due to a reduction in Elevation 120 m asl duration of ice cover; in turn, this is expected to Specific conductivity 39 lS/cm affect limnological variables such as pH, nutrients, pH 8.1 and specific conductivity (Douglas & Smol, 1999). TP 12.7 lg/l These studies suggest that larger water bodies of TN 640 lg/l the high Arctic had a greater thermal inertia which Chl a 1.60 lg/l acted as a buffer against the early-onset of changes DOC 8.2 mg/l in diatom communities (e.g., Doubleday et al., DIC 4.8 mg/l 1995; Michelutti et al., 2003a). NH3 0.011 mg/l Paleolimnological research using abiotic prox- NO2 0.001 mg/l ies, such as varves, have also indicated relatively Cl 1.93 mg/l recent and marked environmental changes in SO4 4.8 mg/l Canadian high Arctic lakes and have similarly SiO2 0.27 mg/l been correlated to climatic warming (Smith et al., POC 0.559 mg/l 2004). Likewise, analyses of sediment cores from 5 PON 0.053 mg/l lakes on Svalbard suggested that climatic warming SRP 0.0057 mg/l was a contributing factor to changing diatom TKN 0.648 mg/l assemblages over the last 150 years (Birks et al., Al 0.02 mg/l 2004; Jones & Birks 2004). Similarly, research Ba 0.0023 mg/l programs in subarctic regions have recorded Cu 0.001 mg/l recent environmental changes consistent with Fe 0.353 mg/l warming (e.g., Sorvari et al., 2002; Ru¨hland et al., Li 0.001 mg/l 2003a). Mn 0.0052 mg/l Despite a growing body of arctic paleoenviron- Sr 0.0112 mg/l mental literature, no detailed paleolimnological inves- Ca 5.2 mg/l tigations have yet been published from the vast western Mg 2.7 mg/l Canadian high Arctic. Instrumental meteorological Na 1.5 mg/l data are from this region are sparse and of short K 0.4 mg/l duration; the nearest weather station is located at Abbreviations are as follows: TP: total phosphorus, unfiltered, Mould Bay, Prince Patrick Island (Fig. 1, site l), TN: total nitrogen, DOC: dissolved organic carbon, DIC: dis- which began collecting data in 1948. This lack of solved inorganic carbon, Chl a: chlorophyll a, POC: particulate instrumental meteorological data hampers our organic carbon, PON: particulate organic nitrogen, SRP: sol- ability to assess past climatic and associated envi- uble reactive phosphorus, TKN: total Kjeldhal nitrogen. ronmental changes from the western Canadian high Arctic. Given the sensitivity of small ponds in the eastern high Arctic (Douglas et al., 1994), a small pond on Melville Island was chosen to be the focus Study site of a high-resolution paleoenvironmental investiga- tion. The goals of this study are to assess diatom- With a surface area of 42 149 km2, Melville Island based paleolimnological changes from a small pond (Fig. 1) is the 4th largest of the Queen Elizabeth on central Melville Island, to examine whether islands, and the 7th largest of all Canadian arctic diatom assemblages changed in composition and, if islands, yet it is uninhabited. Melville Island is the so, when and why these changes occurred. only island in the western high Arctic that retains 18 small permanent ice caps. In the absence of local core from the center of the pond on July 24, 2002. meteorological records, data from Mould Bay The sediment core was sectioned into 0.5 cm (76 13¢ N, 119 19¢ W), Prince Patrick Island intervals from 0 to 13 cm using a Glew (1988) (Fig. 1), located approximately 230 km to the extruder immediately after retrieval. This section northwest, was used to estimate average February of the core was rich in organic matter and con- and July temperature of )34.0 C and 4.0 C, tained mosses. Below 13 cm, the sediment was respectively, and a mean annual precipitation of composed predominantly of fine minerogenic 111.0 mm (Meteorological Service of Canada, sediment, and thus the remainder of the core was 2002). sectioned into 1 cm intervals. Pond MV-AT (unofficial name, 75 19¢ N, Temperature, pH, and specific conductivity 111 25¢ W) is a small, shallow (maximum were measured in the field, and water samples were depth = 0.40 m), alkaline (pH = 8.1), and very sent to the Canadian Center for Inland Waters dilute (specific conductivity = 39 lS/cm) pond (Burlington, Ontario, Canada) for analysis of located on central Melville Island, Northwest nutrients, trace metals, and major ions (Environ- Territories (Fig. 1, site a; Table 1). The pond has ment Canada, 1994). moderately low nutrients (e.g., total phospho- 210Pb dating was performed at the Paleoeco- rus = 12.7 lg/l) and relatively high dissolved logical Environmental Assessment and Research organic carbon (8.2 mg/l) (Table 1). Surficial geol- Lab (PEARL), QueenÕs University, using gamma ogy is composed of a veneer of weathered bedrock spectrometry (Appleby, 2001). Activity levels were (sandstone, siltstone, and shale) of the Griper Bay converted to dates using the Constant Rate of Formation (Hodgson & Vincent, 1984). Pond MV- Supply (CRS) method (Binford, 1990). AT is located on an interfluve approximately 15 m Preparation of diatom samples followed above two braided rivers, with a catchment domi- standard techniques (Battarbee et al., 2001). A nated by grasses, sedges, and mosses. minimum of 300 diatom valves were identified and enumerated at each interval. Diatom identification followed Krammer & Lange-Bertalot (1986–1991), Materials and methods Krammer (2000), Krammer (2002), Lange-Bertalot (2001), and Antoniades (2004). A modified Glew (1989) gravity corer (diame- Diatom results were converted to percent rela- ter = 7.82 cm) was used to collect a 21 cm long tive abundance measures and plotted against

Figure 2. Sedimentation rate for dated sections of the MV-AT core, as calculated based on the Constant Rate of Supply (CRS) method described by Appleby (2001) and Binford (1990). 19 depth using the program C2 version 1.4 beta Results (Juggins, 2003). To identify zones which may be considered to have some statistical validity, opti- Core chronology mal splitting combined with broken-stick analysis was performed (Bennett, 1996), using the program While all 35 sediment intervals were dated, the psimpoll 4.10 (Bennett, 2002). A detrended corre- unsupported 210Pb profile was only detected within spondence analysis (DCA) was applied down-core the upper 7 cm of the sediment core. Gamma to estimate species turnover (Birks, 1998). DCA emissions were low, as is common in high arctic was performed using CANOCO version 4.0 lake sediments (Pienitz et al., 2004), but nonethe- (ter Braak 1998). less the profile exhibited a roughly exponential Loss-on-ignition (LOI) analysis followed Dean decay, indicating that a reliable geochronological (1974). Sediments were freeze-dried and pre- profile could be established for the pondÕs recent weighed before combustion at 550 C for 2 h to history (ca. 200 years). Three intervals had higher provide a proxy for organic matter content of the 210Pb counts relative to the intervals immediately sediment (Heiri et al., 2001). Further combustion above them, and thus were not used in the date at 1000 C for 2 h provided an estimate of calculations; such reversals are not uncommon carbonate content (Dean, 1974). when emissions are very low.

Figure 3. Diatom proﬁle showing the common diatom species found in pond MV-AT. Individual species with >5% relative abundance in at least one sample were retained for the proﬁle; ‘otherÕ is a sum of all other diatoms found in each interval. Dates are based on 210Pb dating using a Constant Rate of Supply model. Percent loss-on-ignition (%LOI 550) is expressed as a % of combustion at 550 C, and is a proxy for organic matter content of the sediment. Percent carbonates (%LOI 1000) is expressed as a percentage of dry weight combusted at 1000 C. Zones are based on optimal splitting and broken-stick analysis. While the most marked species change occurs at 5.25 cm, the shift in diatom assemblage appears to have begun earlier (5.75 cm). 20

Diatoms and LOI

A total of 81 diatom taxa was identified, but only 7 were considered common (i.e., present at >5% relative abundance in at least one interval). Bro- ken-stick analysis based on optimal splitting identified one split of significant note occurring between 5 and 5.5 cm (Fig. 3). From 21 to 6 cm (Zone 1), the diatom assemblage remained relatively constant (Fig. 3), and was dominated by Fragilaria capucina Desmazie` res (40–50%), and secondarily by Achnanthes minutissima Ku¨tzing (20%). Zone 2 begins between 5 and 5.5 cm (Fig. 3). Between 5.5 and 1.0 cm, F. capucina underwent a marked decrease to about 10%, whereas A. minutissima increased to 50% (Fig. 3). Be- tween 3 and 0 cm, Nitzschia perminuta (Grunow) M. Peragallo increased from 10 to 15% relative abundance. At the top of the core, A. minutissima decreased to about 43% relative abundance, with a concordant increase in N. perminuta. Species turnover was low for most of the sediment profile, as the DCA axis 1 species scores remained relatively constant between 1 and 1.25 standard deviation (SD) units from 21 to 5.5 cm (Fig. 4). The total variance of DCA axis 1 was 1.396 SD units. Above 5.5 cm, a sharp decline in the DCA species scores to about 0.25 SD units indicates a strong species shift (Fig. 4). This trend continues until 0.5 cm, at which time there is an inflection in the opposite direction, likely reflecting the resurgence of N. perminuta at the surface of the core (Fig. 4). This shift in direction should be treated with caution, however, since the top of the core contains the least consolidated sediment. The percentage of LOI remained constant at about 17% between 21 and 13 cm, increased Figure 4. Detrended Correspondance Analysis (DCA) of dia- to 20% between 13 and 5.5 cm, and increased tom species scores vs. depth. The DCA axis 1 species scores are further to 50% beginning at 5 cm. (Fig. 3). There scaled in Standard Deviation (SD) units, and provide an esti- was insufficient sediment available to determine a mate of species turnover. value for LOI at the very surface of the core (i.e., between 0 and 0.5 cm). The percent carbonate content remained consistently low (2%) Based on the CRS model, the calculated sedi- throughout the core. mentation rate for the upper 2.5 cm was high (0.0322 cm/yr), relative to other high arctic lakes (e.g., Douglas et al., 1994). Below this level, the Discussion rate dropped off exponentially and between 5 and 6.5 cm, the sedimentation rate reached a plateau at The diatom assemblages from MV-AT remained 0.0106 cm/yr (Fig. 2). stable for most of the pondÕs history captured by 21 this sediment record, with the most striking major species shift predates the onset of artificial N changes occurring early in the 20th century production by at least 30 years (Vitousek et al., (5.25 cm, Fig. 3). This marks the delineation of 1997). In addition, pond MV-AT is currently the only clear zone based on a broken-stick anal- highly P-limited (TN:TP = 50, Table 1). The ysis (Bennett, 1996). As is also evident from the presence of Nostoc balls between 0 and 5 cm in our DCA profile (Fig. 4), a marked shift in species core also suggests that N levels have not been turnover is underway at this time. While these increasing significantly over the 20th century, as changes are most dramatic at this time, it is likely Nostoc (cyanobacteria) are N-fixers and are espe- that the diatom assemblage began to change even cially competitive under low N conditions. Thus, earlier (5.75 cm, Figs. 3 and 4). increases in N would not likely result in the Potential causes contributing to the recent marked diatom changes recorded in this pond. shift in diatom assemblages include atmospheric Instead, these diatom changes are more likely pollution, anthropogenic acidification, artificial related to climatic warming, as discussed below. nitrogen deposition, and/or climatic change. Under conditions of warming, limnological Melville Island is completely uninhabited, and variables such as nutrients, specific conductivity, located far from local pollution sources. While and pH are all expected to increase in a pond such organic pollutants, such as polychlorinated as MV-AT, as a reduction in the duration of ice biphenyls (PCBs), may be transported long dis- cover allows for longer growing seasons and tances (Rose et al., 2004), it is unlikely that the higher primary production (Douglas & Smol, shifts in diatom assemblages recorded here would 1999). While the three most dominant taxa in the be triggered by such contamination. For exam- surface of the core are not known to have highly ple, sites that had direct and prolonged PCB restrictive ecological niches (e.g., Lim et al., 2001b; contamination in Labrador showed no diatom or Michelutti et al., 2003b), autecological informa- chrysophyte response to this input (Paterson et tion gleaned from recent diatom calibration sets in al., 2003). the Arctic and Subarctic (Lim et al., 2001a, b; Diatoms are known to be sensitive indicators of Michelutti et al., 2003b; Ru¨hland et al., 2003b; acidification (e.g., Siver et al., 2003). However, Lim, 2004; Antoniades et al., 2004) suggests that there is no suggestion of acidification of this site as the most likely cause of these assemblage shifts is the current pH is 8.1 and the diatom shift is away environmental changes precipitated by climatic from taxa known to be acidophilic (e.g., Eunotia warming. For example, arctic diatom calibration spp.) and towards species commonly found in sets suggest that A. minutissima has a higher spe- more alkaline arctic waters (Antoniades et al., cific conductivity optimum than F. capucina 2004). The increase in Nitzschia perminuta near the (Antoniades et al., 2004), and is more frequently top of the core provides further evidence that pH found in sites with higher total nitrogen levels was not declining at this site, as N. perminuta has (Lim et al., 2001a; Ru¨hland et al., 2003b) and the highest pH optimum of the three dominant higher dissolved organic carbon (DOC) concen- species in this core, based on a training set from trations (Ru¨hland et al., 2003b). Likewise, N. Ellef Ringnes Island, Arctic Canada (Antoniades perminuta occurs in sites with relatively high pH et al., 2004). and soluble reactive phosphorus levels on Devon Nitrogen deposition has been implicated as a Island (Lim, 2004). Small Nitzschia taxa are also driver of diatom assemblage change in the Rocky often associated with high nutrient loadings in Mountains of Colorado (Wolfe et al., 2001; Saros arctic lakes (e.g., Douglas & Smol, 2000; Michel- et al., 2003; Wolfe et al., 2003). As Achnanthes utti et al., 2002; Douglas et al., 2004). N. permin- minutissima is commonly found in high arctic lakes uta, however, is not commonly associated with any and ponds with relatively high TN values, it is specific environmental variable on Bathurst Island possible that some degree of anthropogenic nitro- (Fig. 1, site m; Lim et al., 2001a; Lim et al., gen deposition may be affecting MV-AT as well. It 2001b), the central Canadian Subarctic (Ru¨hland is unlikely, however, that this deposition could et al., 2003b), or Ellef Ringnes Island (Antoniades have initiated the shift in diatom communities on et al., 2004). As algal production is expected to Melville Island because the commencement of the increase with a lengthening of the growing season, 22 a shift toward diatom species that have affinities for higher nutrient, DOC, and specific conductivity levels suggests that the environment experienced a warming trend at MV-AT beginning in the early 20th century. The other common species found in Figure 3 occur in very low abundances. They are comprised by three Cymbella and one Eunotia spp. It is likely that the decrease in Eunotia further supports a rising pH over time in this site, as this genus is characteristic of acidic waters. Cymbella spp. are commonly associated with shallow water arctic habitats (Douglas & Smol, 1995; Michelutti et al., 2003b). Furthermore, it should be noted that all of the 7 common taxa are periphytic (Fig. 3). F. capucina is known to be commonly associated with rocks, moss, and sediment on Victoria Island (Fig. 1, site n; Michelutti et al., 2003b), while N. perminuta and A. minutissima are both commonly found on moss, rocks, and sediment on Victoria Island (Michelutti et al., 2003b), Cape Hershel (Douglas & Smol, 1995), and Bathurst Island (Lim et al., 2001b). Thus, the changes in the diatom assemblage do not represent a shift between planktonic and benthic taxa. Concurrent with the marked floristic shifts in this core, there was also a sharp rise in %LOI near the 5 cm depth (Fig. 3). Increased run-off from the catchment may result in a greater amount of allochthonous organic matter washing-in to the pond. However, evidence for enhanced autochthonous production is apparent in the correspondence between high %LOI values of 30–50% and the presence of many preserved Nostoc (cyanobacteria) balls in the sediment core (up until 5 cm depth). The occurrence of Nostoc may be a further indicator of climatic warming, as increased temperatures would lengthen the growing season.

Comparison to other paleoenvironmental records Figure 5. Mean June–July–August temperature data and an- The diatom proﬁle from MV-AT exhibits species nual precipitation data from Mould Bay, Prince Patrick Island (see Fig. 1 for location) between 1948 and 1996 (Meteorological shifts that are consistent with recent environmental Service of Canada, 2004). The smoothed line is a LOWESS change inferred from diatom-based paleolimno- curve with a span of 0.35. logical records from other areas in the Canadian high Arctic (Fig. 1, Douglas et al., 1994; Double- prior to 1919, the more recent MV-AT commu- day et al., 1995; Perren et al., 2003; Michelutti et al., nities share similarities to other sites. For example, 2003a; Antoniades et al., 2005). While none of the pond I-F (Isachsen, Ellef Ringnes Island) recorded previously recorded diatom proﬁles show similar increases in Nitzschia perminuta since 1850 A.D. assemblages to those dominating the MV-AT core (Antoniades et al., 2005). In Self Pond (Alert, 23

Ellesmere Island), the increased relative abundance Mean June–July–August temperature data indi- of Achnanthes minutissima tracked increases of cate that there is little trend in temperature both pH and warming temperatures since 1920 between 1948 and 1996 (Fig. 5, Meteorological A.D. (Antoniades et al., 2005). The similarities Service of Canada, 2004), likewise, the diatom between these modern assemblages and those record does not show any major shifts after found in MV-AT provide further evidence that the 1950. The increase in Nitzschia perminuta near diatom changes in MV-AT may indicate recent the top of the core (Fig. 3) is not captured in the warming. Furthermore, the assemblages in other meteorological data, which ends in 1996. There studies are different, but the nature of the diatoms does appear to be an increase in annual precipi- shifts are similar to that of MV-AT, and have been tation values at Mould Bay (Fig. 5, Meteorologi- interpreted as a response to warming (e.g., Douglas cal Service of Canada, 2004), but this is not et al., 1994; Doubleday et al., 1995; Gajewski et al., reflected in the diatom profile of MV-AT. A closer 1997; Perren et al., 2003). The onsets of these dia- examination of the precipitation values for June, tom shifts, however, have occurred at different July, and August (the months representing and times. The large diatom assemblage shift in shouldering the open water season) indicate only a MV-AT appears to have started 1919, which very slight increase in precipitation since 1948 at would be consistent with the timing of diatom Mould Bay (data not shown). changes at both Sawtooth Lake (Fosheim Penin- Other abiotic proxy environmental records sula, Ellesmere Island, Fig. 1, site e; Perren et al., from the Canadian high Arctic also show 20th 2003), and Self Pond (Antoniades et al., 2005), but century warming. For example, varved sediment later than diatom assemblage changes at ponds records from Tuborg Lake, Ellesmere Island near Isachsen (Antoniades et al., 2005), and at (Fig. 1, site d), were interpreted to show warming Cape Herschel, Ellesmere Island (Douglas et al., beginning after 1865, and especially after 1908 1994). Assuming the dating of each of the sediment (Smith et al., 2004). Likewise, ice core melt re- profiles is reasonably accurate, the discrepancy in cords from the Devon Ice Cap (Fig. 1, site f) timing may be due to a regional difference in suggest warm temperatures since 1869, and par- warming in the western Canadian high Arctic. ticularly after 1925 (Koerner, 1977). Agassiz Ice The only available paleoenvironmental record Cap, Ellesmere Island (Fig. 1, site h), melt layers for Melville Island is the glacier mass balance data also suggested that the 20th century has been the (Koerner, 2002) from the Melville South Ice Cap warmest of the last millenium (Koerner & Fisher, (Fig. 1, site i). Glaciers cover 160 km2 of Mel- 1990), and analysis of ice fabric, dirt and firn ville Island (Ommanney, 2002). Mass balance from Meighen Island ice cap (Fig. 1, site j) sug- measurements from the Melville South Ice Cap gested warm conditions between 1884 and 1964 have been made annually (with some exceptions) (the extent of the record) (Koerner & Paterson, by the Geological Survey of Canada since 1964 1974). These records, although all located sig- (Ommanney, 2002). Although these records show nificant distances from Melville Island (Fig. 1), all that there is little trend in the mass balance data, suggest that regional climatic warming accelerated each year since 1968 (with the exception of 1984, into the 20th century. The dramatic diatom 1986, and 1991) has had an average negative mass change at MV-AT appears to be consistent, in balance, indicating that melt had exceeded accu- both time and nature, to these paleoenviron- mulation (Koerner, 2002). This is consistent with mental records. the trend inferred from our diatom and %LOI data. The nearest available long-term meteorological Conclusion data are from Mould Bay, Prince Patrick Island (Figs. 1 and 5). Although this record extends back Diatom species assemblages have changed mark- to only 1948, and hence does not capture the edly since the early 20th century in this small pond inception of the large diatom and %LOI shifts on Melville Island, suggesting higher pH, specific recorded here, it may still be used to assess more conductivity, and nutrient levels. These data, recent temperature and precipitation variations. coupled with increases in % LOI (as a proxy for 24 organic matter), imply that algal production in- Antoniades, D., M. S. V. Douglas & J. P. Smol, 2005. Quan- creased beginning around 1919, and are consistent titative estimates of recent environmental changes in the with environmental changes projected under a Canadian High Arctic inferred from diatoms in lake and pond sediments. Journal of Paleolimnology 33: 349–360. 20th century climate warming scenario. The Appleby, P. G., 2001. Chronostratigraphic techniques in recent direction of environmental change in pond MV- sediments. In Last, W. M. & J. P. Smol (eds), Tracking AT is broadly similar to that found in both lakes Environmental Change Using Lake Sediments Volume 1: and ponds throughout the eastern Canadian high Basin Analysis, Coring, and Chronological Techniques. Arctic, the Canadian Subarctic, and Svalbard. The Kluwer Academic Publishers, Dordrecht: 171–203. Battarbee, R. W., V. J. Jones, R. J. Flower, N. G. Cameron, H. onset of these environmental changes, however, is Bennion, L. Carvalho & S. Juggins, 2001. Diatoms. In Smol, somewhat later than those of small ponds from J. P., H. J. B. Birks, & W. M. Last (eds), Tracking Envi- eastern Ellesmere Island (Douglas et al., 1994), but ronmental Change Using Lake Sediments: Terrestrial, Algal, similarly timed to changes in Self Pond (Antoni- and Siliceous Indicators. Kluwer Academic Publishers, ades et al., 2005) and Sawtooth Lake (Perren et al., Dordrecht: 155–202. Bennett, K. D., 1996. Determination of the number of zones in a 2003). As expected, the changes in diatoms from biostratigraphical sequence. New Phytologist 132: 155–170. MV-AT occur much earlier and more dramatically Bennett, K. D., 2002. Documentation for psimpoll 4.10 and than the shifts apparent in relatively large lakes pscomb 1.03. Department of Earth Sciences. Uppsala Uni- from the eastern Canadian high Arctic (e.g., versitet, Sweden. Michelutti et al., 2003a). Future paleolimnological Binford, M. W., 1990. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake sediment cores. Journal studies in this region will help to refine and of Paleolimnology 3: 253–267. corroborate these interpretations. Birks, H. J. B., 1998. Numerical tools in palaeolimnology – Progress, potentialities, and problems. Journal of Paleolim- nology 20: 307–332. Acknowledgements Birks, H. J. B., V. J. Jones & N. L. Rose, 2004. Recent environmental change and atmospheric contamination on Sval- bard as recorded in lake sediments – synthesis and general This research was funded by Natural Sciences conclusions. Journal of Paleolimnology 31: 531–546. and Engineering Research Council (NSERC) Dean, W. E. Jr., 1974. Determination of carbonate and organic grants to BEK, MSVD, and JPS, and a Northern matter in calcareous sediments and sedimentary rocks by Scientific Training Program grant and a QueenÕs loss on ignition: comparison with other methods. Journal of Sediment Petrology 44: 242–248. Graduate Award to BEK. Logistical support was Doubleday, N. C., M. S. V. Douglas & J. P. Smol, 1995. Pa- provided by the Polar Continental Shelf Project leoenvironmental studies of black carbon deposition in the (PCSP). Assistance in the field was provided by High Arctic – a case-study from northern Ellesmere Island. J.R. Glew, N. Michelutti, and especially by Science of the Total Environment 161: 661–668. D. Antoniades. Input from three reviewers Douglas, M. S. V. & J. P. Smol, 1995. Periphytic diatom assemblages from high arctic ponds. Journal of Phycology 31: 60–69. greatly improved this manuscript. Many thanks Douglas, M. S. & V. J. P. Smol, 1999. Freshwater diatoms as to K. Ru¨hland, D. Selbie, A. Harris, A. Strecker, indicators of environmental change in the High Arctic. In and R. Bull for helpful comments. This is PCSP Stoermer, E. F. & J. P. Smol (eds), The Diatoms: Applica- contribution #05004. tions for the Environmental and Earth Sciences. Cambridge University Press, Cambridge: 227–244. Douglas, M. S. V. & J. P. Smol, 2000. Eutrophication and recovery in the High Arctic: Meretta lake (Cornwallis Island, References Nunavut, Canada) revisited. Hydrobiologia 431: 193–204. Douglas, M. S. V., J. P. Smol & W. Blake Jr, 1994. Marked Antoniades, D., 2004. The limnology, diatom autecology, and post-18th century environmental change in high-arctic eco- paleolimnology of lakes and ponds from Alert, Ellesmere systems. Science 266: 416–419. Island, Isachsen, Ellef Ringnes Island, and Mould Bay, Douglas, M. S. V., J. P. Smol, J. M. Savelle & J. M. Blais, 2004. Prince Patrick Island, Canadian High Arctic. PhD Thesis, Prehistoric Inuit whalers affected arctic freshwater ecosys- University of Toronto. tems. Proceedings of the National Academy of Sciences of Antoniades, D., M. S. V. Douglas & J. P. Smol, 2004. Diatom the United States of America 101: 1613–1617. species–environment relationships and inference models Environment Canada, 1994. Manual of Analytical Methods – from Isachsen, Ellef Ringnes Island, Canadian High Arctic. National Laboratory for Environmental Testing, Canadian Hydrobiologia 529: 1–18. Centre for Inland Waters. Burlington, ON. 25

Fisher, D. A., 1979. Comparison of 105 years of oxygen Krammer, K., 2002. Diatoms of Europe: Cymbella. A.R.G. isotope and insoluble impurity profiles from the Devon Gantner Verlag, Ruggell. Island and Camp Century ice cores. Quaternary Research Krammer, K. & H. Lange-Bertalot, 1986–1991. Bacillariophy- 11: 299–305. ceae. In Ettl, H., J. Gerloff, H. Heynig & D. Mollenhauer Gajewski, K., P. B. Hamilton & R. McNeely, 1997. A high (eds), Gustav Fischer Verlag, Stuttgart. resolution proxy-climate record from an arctic lake with Lange-Bertalot, H., 2001. Diatoms of Europe. Navicula sensu annually-laminated sediments on Devon Island, Nunavut, stricto, 10 Genera Separated from Navicula sensu lato, Canada. Journal of Paleolimnology 17: 215–225. Frustulia. A.R.G. Gantner Verlag, Ruggell. Glew, J. R., 1988. A portable extruding device for close interval Lim, D. S. S., 2004. Limnology and diatom paleoecology of sectioning of unconsolidated core samples. Journal of Pa- diatoms on Banks Island, N.W.T., and Devon Island, Nun- leolimnology 2: 241–243. avut, Canadian Arctic, PhD Thesis. University of Toronto. Glew, J. R., 1989. A new trigger mechanism for sediment Lim, D. S. S., M. S. V. Douglas & J. P. Smol, 2001a. Diatoms samplers. Journal of Paleolimnology 2: 241–243. and thier relationship to environmental variables from lakes Grumet, N. S., C. P. Wake, P. A. Mayewski, G. A. Zielinski, S. and ponds on Bathurst Island, Nunavut, Canadian High I. Whitlow, R. M. Koerner, D. A. Fisher & J. M. Woollett, Arctic. Hydrobiologia 450: 215–230. 2001. Variability of sea-ice extent in Baffin Bay over the last Lim, D. S. S., C. Kwan & M. S. V. Douglas, 2001b. Periphytic millenium. Climatic Change 49: 129–145. diatom assemblages from Bathurst Island, Nunavut, Cana- Heiri, O., A. F. Lotter & G. Lemcke, 2001. Loss on ignition as a dian High Arctic: An examination of community relation- method for estimating organic and carbonate content in ships and habitat preferences. Journal of Phycology 37: sediments: reproducibility and comparability of results. 379–392. Journal of Paleolimnology 25: 101–110. Meteorological Service of Canada, 2002. Canadian Climate Hodgson, D.A. & J.-S. Vincent, 1984. Surficial geology, Central Normals 1971–2000. Environment Canada, Ottawa. Melville Island, Northwest Territories; Geological Survey of Meteorological Service of Canada 2004. Climate Data Online. Canada, Map 1583A, scale 1:250 000. Environment Canada, Ottawa. http://www.climate.weather- Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van der office.ec.gc.ca/climateData/canada_e.html, accessed Decem- Linden & D. e. Xiaosu, 2001. Climate Change 2001: The Sci- ber 10, 2004. entific Basis Contribution of Working Group I to the Third Michelutti, N., M. S. V. Douglas & J. P. Smol, 2002. Tracking Assessment Report of the Intergovernmental Panel on Climate recent recovery from eutrophication in a high arctic lake Change (IPCC). Cambridge University Press, Cambridge. (Meretta Lake, Cornwallis Island, Nunavut, Canada) using Jones, V. J. & H. J. B. Birks, 2004. Lake-sediment records of fossil diatom assemblages. Journal of Paleolimnology 28: recent environmental change on Svalbard: results of diatom 377–381. analysis. Journal of Paleolimnology 31: 445–466. Michelutti, N., M. S. V. Douglas & J. P. Smol, 2003a. Diatom Juggins, S., 2003. C2: Software for Ecological and Palaeoco- response to recent climatic change in a high arctic lake (Char logical Data Analysis and Visualisation, version 1.3. Lake, Cornwallis Island, Nunavut). Global and Planetary Department of Geography, University of Newcastle. Change 38: 257–271. Kaufman, D. S., T. A. Ager, N. J. Anderson, P. M. Anderson, Michelutti, N., A. J. Holtham, M. S. V. Douglas & J. P. Smol, J. T. Andrews, P. J. Bartlein, L. B. Brubaker, L. L. Coats, L. 2003b. Periphytic diatom assemblages from ultra-oligo- C. Cwynar, M. L. Duvall, A. S. Dyke, M. E. Edwards, W. R. trophic and UV transparent lakes and ponds on Victoria Eisner, K. Gajewski, A. Geirsdo´ttir, F. S. Hu, A. E. Jen- Island and comparisons with other diatom surveys in the nings, M. R. Kaplan, M. W. Kerwin, A. V. Lozhkin, G. M. Canadian Arctic. Journal of Phycology 39: 465–480. MacDonald, G. H. Miller, C. J. Mock, W. W. Oswald, B. L. O’Brien, S. R., P. A. Mayewski, L. D. Meeker, D. A. Meese, M. Otto-Bliesner, D. F. Porinchu, K. Ru¨hland, J. P. Smol, E. J. S. Twickler & S. I. Whitlow, 1995. Complexity of Holocene Steig & B. B. Wolfe, 2004. Holocene thermal maximum in climate as reconstructed from a Greenland ice core. Science the western Arctic (0–180 W). Quaternary Science Reviews 270: 1962–1964. 23: 529–560. Ommanney, C. S. L., 2002. Glaciers of Canada. In R. S. Wil- Koerner, R. M., 1977. Devon Island Ice Cap: core stratigraphy liams & J. G. Ferrigno (eds), Glaciers of North America. U.S. and paleoclimate. Science 196: 15–18. Geological Survey Professional Paper 1386-J-1, J27–J82. Koerner, R. M., 2002. Glaciers of high arctic islands. In R. S. Paterson, A. M., A. A. Betts-Piper, J. P. Smol & B. A. Zeeb, Williams & J. G. Ferrigno (eds), Glaciers of North America. 2003. Diatom and chrysophyte algal response to long-term U.S. Geological Survey Professional Paper 1386-J-1, PCB contamination from a point-source in northern Lab- J111–J146. rador, Canada. Water Air and Soil Pollution 145: 377–393. Koerner Fisher, R. M. D. A., 1990. A record of Holocene Paterson, W. S. B., R. M. Koerner, D. A. Fisher, S. Johnsen, summer climate from a Canadian high-Arctic ice core. H. B. Clausen, W. Dansgaard, P. Bucher & H. Oeschger, Nature 343: 630–631. 1977. An oxygen-isotope climate record from the Devon Koerner, R. M. & W. S. B. Paterson, 1974. Analysis of a core Island ice cap, arctic Canada. Nature 266: 508–511. through the Meighen Ice Cap, Arctic Canada, and its pa- Perren, B. B., R. S. Bradley & P. Francus, 2003. Rapid lacus- leoclimatic implications. Quaternary Research 4: 253–263. trine response to recent High Arctic warming: A diatom Krammer, K., 2000. Diatoms of Europe: The Genus Pinnu- record from Sawtooth Lake, Ellesmere Island, Nunavut. laria. A.R.G. Gantner Verlag, Ruggell. Arctic Antarctic and Alpine Research 35: 271–278. 26

Pienitz, R., M. S. V. Dougas & J. P. Smol (eds), 2004. Long- Smol, J. P., 2002. Pollution of Lakes and Rivers: A Paleoen- Term Environmental Change in Arctic and Antarctic Lakes. vironmental Perspective. Arnold, London. Springer, Dordrecht. Sorvari, S., A. Korhola & R. Thompson, 2002. Lake diatom Rose, N. L., C. L. Rose, J. F. Boyle & P. G. Appleby, 2004. response to recent Arctic warming in Finnish Lapland. Lake-sediment evidence for local and remote sources of Global Change Biology 8: 171–181. atmospherically deposited pollutants on Svalbard. Journal Stoermer, E. F. & J. P. Smol, 1999. The Diatoms: Applications of Paleolimnology 31: 499–513. for the Environmental and Earth Sciences. Cambridge Ru¨hland, K., A. Priesnitz & J. P. Smol, 2003a. Paleolimno- University Press, Cambridge. logical evidence from diatoms for recent environmental ter Braak, C. J. F., 1998. CANOCO and CANOPost, Version 4 changes in 50 lakes across Canadian arctic treeline. Arctic for Windows. Microcomputer Power, Ithaca, NY. Antarctic and Alpine Research 35: 110–123. Vitousek, P. M., J. D. Aber, R. W. Howarth, G. E. Likens, Ru¨hland, K. M., J. P. Smol & R. Pienitz, 2003b. Ecology and P. A. Matson, D. W. Schindler, W. H. Schlesinger & D. G. spatial distributions of surface-sediment diatoms from 77 Tilman, 1997. Human alteration of the global nitrogen cycle. lakes in the subarctic Canadian treeline region. Canadian Ecological Applications 7: 737–750. Journal of Botany 81: 57–73. Wolfe, A. P., 2000. A 6500-year record from southwestern Saros, J. E., S. J. Interlandi, A. P. Wolfe & D. R. Engstrom, Fosheim Peninsula, Ellesmere Island, Nunavut. In Garneau 2003. Recent changes in the diatom community structure of Alt, M. B. T. (ed.) Environmental Response to Climate lakes in the Beartooth Mountain Range, U.S.A. Arctic, Change in the Canadian High Arctic. Geological Survey of Antarctic, and Alpine Research 35: 18–23. Canada, Ottawa: 249–256. Serreze, M. C., J. E. Walsh, F. S. Chapin III, T. Osterkamp, Wolfe, A. P., J. S. Baron & R. J. Cornett, 2001. Anthropogenic M. Dyurgerov, V. Romanovsky, W. C. Oechel, J. Morison & nitrogen deposition induces rapid ecological changes in al- T. R. G. Zhang Barry, 2000. Observational evidence of pine lakes of the Colorado Front Range (USA). Journal of recent change in the northern high-latitude environment. Paleolimnology 25: 1–7. Climatic Change 46: 159–207. Wolfe, A. P., A. C. Gorp & J. S. Baron, 2003. Recent eco- Siver, P. A., R. Ricard, R. Goodwin & A. E. Giblin, 2003. logical and biogeochemical changes in alpine lakes of Estimating historical in-lake alkalinity generation from sul- Rocky Mountain National Park (Colorado, USA): a re- fate reduction and its relationship to lake chemistry as in- sponse to anthropogenic nitrogen deposition. Geobiology 1: ferred from algal microfossils. Journal of Paleolimnology 29: 153–168. 179–197. Zdanowicz, C. M., G. A. Zielinski, C. P. Wake, D. A. Fisher & Smith, S. V. Jr., R. S. Bradley & M. B. Abbott, 2004. A 300 R. M. Koerner, 2000. A Holocene record of atmospheric year record of environmental change from Lake Tuborg, dust deposition on the Penny Ice Cap, Baﬃn Island, Canada. Ellesmere Island, Nunavut, Canada. Journal of Paleolim- Quaternary Research 53: 62–69. nology 32: 137–148.