Reply to comment by K. Gajewski on “Abrupt environmental change in Canada's northernmost lake”



[1] We welcome the opportunity to respond to Gajewski's [2008] comments on our study [Antoniades et al., 2007, hereinafter referred to as A07]. We demonstrate here that his assertions are not supported by the evidence and that the conclusions in A07 are based on a sound interpretation of the data.

1. Sample Collection

[2] Sediment coring by scientific divers is one of several effective approaches to core collection, each capable of intact sediment recovery. Divers sampling in situ can observe the sediment surface and avoid sampling disturbed sediments, something not guaranteed by conventional methods. Our scientific divers are the world's most experienced at diving with scuba beneath polar lake ice, with extensive experience sampling the bottom of lakes while preserving sediment integrity [Andersen, 2007]. The sediments recovered were completely undisturbed by divers or ice (D. Andersen, personal observation, 2003).

[3] Gajewski's comments about frozen sediments are not applicable to our study site, Ward Hunt Lake (WHL). Given the description of conditions “beneath the thick ice cover” and the need for divers, it was implicit that our core was not from the lake's shallow moat zone. The core was taken along the centre line of the lake at ∼5m depth, under thick perennial ice in a water layer 1–1.5m thick, and well away from the moat. Numerous observations at this site indicate that liquid water persists in WHL throughout the year and that the sediments do not freeze in winter. Our divers also reported that no bottom ice was present. Although Gajewski notes (citing Nichols [1967]) the problems inherent in working with frozen sediment, our sediments were not frozen. We carefully inspected the core for signs of disturbance and there were none. There was no deformation from freezing in the upper visible strata or anywhere else. The preservation in the core of visible layers and marked shifts of pigments, diatoms and organic matter content reinforced our confidence in the record's integrity.

2. Dating of Sediments

[4] Obtaining a reliable chronology is often the biggest challenge in Arctic paleolimnology, and our study was no exception. However, Gajewski's assertion that our core “is essentially undated” is unwarranted. Sediments with undetectable excess 210Pb and 137Cs are frequently encountered in the polar regions [Wolfe et al., 2004]. A07 reported a weak 137Cs peak, and indicated the upper sediments could not be constrained beyond our cautious inference “that the upper 0.5 cm of sediment contain the last 50 years of sediment deposition” [A07; p. 3].

[5] Gajewski judges our radiocarbon date to be “extremely suspect”, yet provides no justification. Although 14C dating of sediment must be approached with caution [Wolfe et al., 2004], Arctic studies have shown that sediment can be used to produce accurate, unbiased dates [e.g., Peros and Gajewski, 2008]. There was no identifiable organic material in the core for dating, and to be conservative, when calculating error ranges we sought to “encompass the maximum possible range of uncertainty given the age of the core” [A07; p.3]. Roughly two thirds of WHL's catchment is underlain by syenite and schist, with the remaining third underlain by sedimentary rocks, including carbonates. There is no coal in the area. Prior to radiocarbon analysis, the sediment sample was crushed and treated with HCl to remove all carbonates. Although we were conscious of the possibility of old carbon contamination, there is no evidence for its occurrence. In polar lakes, such contamination is often due to inorganic carbon inputs from glacial meltwater [Doran et al., 1999]; there are no glaciers in WHL's catchment. The sediment δ13C (Table 1) closely approximates those measured from nearby microbial mats on the Ward Hunt Ice Shelf that are not susceptible to hardwater effects [Mueller and Vincent, 2006]. Benthic mats dominate carbon fixation in WHL [Bonilla et al., 2005], and have been present throughout the lake's history [A07]. Similar microbial mats in Antarctic lakes were shown to be equilibrated with modern 14CO2 [Doran et al., 1999]. Low and reasonably constant pigment concentrations throughout most of the WHL core attest to a nearly frozen lake with limited productivity in its moat. The carbon source for our radiocarbon date is consistent with moat productivity in equilibrium with atmospheric 14CO2. While our 14C date implies slower sedimentation rates below the 137Cs-dated horizon, such changes are common in Arctic cores, where they result from concurrent increases in productivity and sedimentation, and autocompaction of deeper sediments. Polar lakes preserve surprisingly long records within short sequences, and short cores can often contain the entire Holocene [Wolfe et al., 2004]. There were neither lithological changes nor changes in our diagenetic index suggesting changes in preservation conditions or hiatuses. A large volume of evidence thus supports the reliability of our 14C age, and there is an absence of any evidence for its rejection.

Table 1. 14C Dating Information for the Sediment Core From Ward Hunt Lake
IntervalMaterialConventional 14C Age (years B.P.)δ13C (‰ PDB)Calibrated Age (cal years B.P.)2-sigma Calibrated RangeaLab Code
  • a

    The 2-sigma calibrated range represents the 95% confidence interval for the age of the sample.

15–16 cmBulk sediment7760 ± 40−22.784508299–8601Beta-229887

3. Pigment Data

[6] Gajewski provides no specific criticisms of our pigment record, and instead casts aspersions on the technique. However, HPLC pigment analysis is a well-established tool for paleoecological inferences [Leavitt and Hodgson, 2001] and climate reconstructions in the polar regions [e.g., Hodgson et al., 2006]. Our index of pigment diagenesis [A07, Figure 2g] addressed the issue of changing diagenetic conditions within the core. The index's relative stability across the horizon of pigment increase supports the conclusion that this increase reflected an ecological shift in WHL, and not diagenesis. Moreover, our HPLC pigment trajectories mirror those indicating recent productivity increases in Arctic lakes unprecedented in 5000 years, measured by reflectance spectroscopy [Michelutti et al., 2005].

[7] The instrument error of our HPLC (expressed as the coefficient of variation in %, CV) for chlorophylls with concentrations similar to chl-a in our core was 1.0–3.2%, while for our least concentrated chlorophyll (chl-b), the CV was 1.9–8.1%. For all carotenoids, the CV was 10.9–14.4% at the low concentrations that were present. We have subsequently run triplicate samples from the WHL core. The CVs for individual pigments above the horizon of pigment increase were 4–14% (mean = 10%), while at the low concentrations present below this horizon they were 2–35% (mean = 22%). However, the absolute concentrations of almost all major pigments increased in our core by three orders of magnitude, placing the measured changes far outside instrument and replication error.

4. Diatom Analysis

[8] We are familiar with the effects of diatom dissolution. As discussed in A07, we specifically looked for signs of diatom degradation in the WHL record and found none. High latitude sediments can preserve diatoms and other proxy indicators on scales of tens to hundreds of thousands of years without dissolution [Hodgson et al., 2006; Wolfe et al., 2000]. Diatom paleoproductivity may be inferred from absolute diatom abundances (as per Gajewski's suggestion) or from pigments, which include general productivity indicators (i.e., chlorophyll-a, β-carotene) and those produced by diatoms (i.e., diadinoxanthin, fucoxanthin). This information was already presented [A07, Figure 2]. The synchronicity between the large pigment increases and the appearance of diatoms provide independent, compelling evidence for abrupt ecological change in WHL.

[9] Gajewski's four alternate conclusions are untenable, given the following:

[10] 1. We examined the core for evidence of disturbance, and there was none.

[11] 2. Sedimentary and microbial mat δ13C values support a 14C date from material consistent with moat productivity in equilibrium with atmospheric 14CO2.

[12] 3. Diatom concentrations were reported in A07 as pigment proxies of diatom productivity. We looked for evidence of diatom dissolution and found none.

[13] 4. Our pigment analyses and index of diagenesis refute the hypothesis that these changes were diagenetic.

[14] The synoptic reasons that Gajewski seeks for WHL's relative insensitivity to climate are simple: it is a small lake, on a small island surrounded by the Northern Hemisphere's largest ice shelf. The lake is the northernmost in North America, in one of the world's most ice-dominated regions. These characteristics result in a microclimate cooler than elsewhere in the Arctic. Our data from the Ward Hunt Island weather station show that monthly mean temperatures are correlated with nearby Alert [r2 = 0.991, p < 0.0001], and that summer mean temperature is an average of 1.3°C colder than Alert. The relatively cooler microclimate is confirmed by WHL's perennial, 4m thick ice cover while the overwhelming majority of Arctic lakes, including those at Alert, have only seasonal ice.

[15] Contrary to Gajewski's assertion, we have no “preconceived notions of Holocene climate variability”. We judged our sedimentary record based on its own merits. A07 made no statement about spatial variability in Holocene Arctic climates, and drew no conclusions about latitudinal amplification of climate change or any aspect of human-induced global warming. In this regard Gajewski appears to have confused our article with others, including Smol et al. [2005], who addressed some of these issues. A07 concluded, based on the data at hand, that during the Holocene, dramatic ecological responses in WHL were elicited recently by unprecedented warming.

[16] Among studies that Gajewski cites that “show a warm early Holocene”, his own “provides…little information about the middle and early Holocene” [Finkelstein and Gajewski, 2007, p. 808]. Smith [2002, Figure 9] suggests “cooler” conditions prior to ∼4000 BP, and notes that his inferences cannot be compared to recent changes, since the upper sediments were lost during coring. Several others [Hyvärinen, 1985; Gajewski, 1995; Peros and Gajewski, 2008] are based solely on pollen data. Due to limited local production and the frequent domination of long-range pollen deposition, “limits to the interpretation of past environments are particularly significant” in Arctic lakes [Gajewski et al., 1995, p. 609]. Pertinent Arctic studies not cited by Gajewski, yet with secure 14C or varve chronologies, show marked recent changes following several millennia of ecological stability [e.g., Douglas et al., 1994; Perren et al., 2003; Michelutti et al., 2005].

[17] It is not our intent to debate the relative merits of the Arctic paleolimnological literature, but rather to suggest that it provides a less homogenous picture of Holocene Arctic climate than Gajewski implies. The Canadian Arctic Archipelago encompasses roughly 1.42 million km2, with pronounced gradients in climatic and limnological conditions. There is no reason to expect that either climate changes or their ecological effects were uniform across this expanse, and Arctic climate has varied disparately during the Holocene [Kaufman et al., 2004].While the bulk of Arctic paleolimnological literature clearly shows that recent warming has had more profound ecological effects than other prior Holocene changes, this does not discount Gajewski's suggestion of warm conditions during the early Holocene. Earlier warm periods, if they occurred at WHL, may not be recorded in the sediments because they were insufficient to drive the lake across critical ecological thresholds, while recent environmental changes have clearly done so. Regardless of the exact spatial and temporal variations in Holocene Arctic climate, A07 established unambiguously that pronounced and unprecedented ecological shifts recently occurred in Canada's northernmost lake.