Field parameters and major ions
The results of the hydrochemistry and environmental isotopes for the 40 lakes are presented spatially in Figs. S1–S11 and are located in Tables S1 and S2.
The lake waters are oxic (8.6–12.6 mg l−1) and range from slightly acidic (pH 6.0) to slightly alkaline (pH 9.2). Lake water temperatures are generally highest for lakes along the west coast (greater than 10 °C, Table S2). Phosphate concentrations are below detection level (0.1 mg l−1) for all lakes and nitrate was low ranging from below detection limit (< 0.05) to 0.21 mg l−1. A number of lakes have similar ionic ratios to seawater with Na–Cl type waters being the dominant cation and anion. Bicarbonate concentrations were calculated by difference and were low with a maximum of 36 mg l−1 and average of 3.7 mg l−1. All lakes are low in Cl concentrations ranging from 1.4 (LK22) to 3.7 (LK3) mmol l−1 and Na concentrations ranging from 0.99 (LK28) to 2.57 (LK3) mmol L−1. The SO4, Cl, Mg and Na concentrations follow a similar pattern to seawater in the Schoeller plot (Fig. 2), with all showing very high significant positive correlations (⍴ ≥ 0.75, p ≤ 1.4 × 10–5, Table S3). The K, Ca, F and Si concentrations follow a similar pattern to seawater for some lakes, whilst others diverge, with all lake waters containing higher Si concentrations than seawater (Fig. 2). An increase in Cl concentration is broadly reflected in the increase in Br, K and Sr (all ⍴ ≥ 0.71, p ≤ 1.7 × 10–6). The high correlation between these variables implies a similar source of ions, or that the waters have undergone similar hydrochemical processes. Some variables such as F and Cl, F and K, and Cl and Si, however, are not significantly correlated (⍴ = 0.24, − 0.26 and − 0.14, p = 0.1, 0.1 and 0.4 respectively, Table S3) suggesting they have a different source.
Lake waters sampled from Macquarie Island represented as a Scholler plot with the red dashed bolded line representing seawater composition and black lines are lake water concentration of SO4, Cl, Mg, Na, K, Ca, F and Si.
Hierarchical cluster and principal component analysis
Hierarchical cluster analysis revealed three main clusters of lakes (Fig. 3). These clusters were used to colour code the results of the PCA to determine the main hydrochemical processes for the groupings. Cluster 1 represents lakes containing high SO4, Cl, Na, EC, Br, Sr, K, Mg, δ2H, Al and DOC concentration, and low pH, δ13CDIC and smaller distance from the west coast (all p < 0.5, Table S4). Cluster 2 represents lakes containing high Ca, F, SiO2, d-excess, pH and Fe, low elevations, K and δ2H (all p < 0.5, Table S4). Cluster 3 represents lakes containing low Na, EC, F, Cl, Ca, SO4, Fe, Mg, DOC concentrations, Br, SiO2, Al and are located at higher elevations (all p < 0.5, Table S4).
(a) Hierarchical cluster analysis of lake water chemistry variables on Macquarie Island, showing three lake clusters and the five most significant variables associated with each cluster (Table S4). (b) Lakes coloured by cluster membership superimposed on a layer of the islands elevational variation.
PCA results show two main components explaining a total of 63% of the variability in the dataset. Components 1 and 2 of the PCA explain 41.9% and 21.1%, respectively. Variables loading most strongly on the first component (Dim1) include K, SO4, Sr, d-excess, Cl, δ18O, δ2H, Na and Br (Fig. S12). Variables loading most strongly on the second component (Dim2) include Ca, F, Fe, SiO2, DOC concentration, EC, elevation, Na and Mg (Fig. S12). Lakes on the western side of the island strongly influence Dim1 (i.e. LK3, LK6, LK10, LK23, LK30, LK37 and LK38) whilst lakes on the central and eastern portion of the island (i.e. LK11, LK12, LK21, LK24, LK40 and LK42) influenced Dim2 most strongly.
Environmental isotopes and dissolved organic carbon concentrations
The δ2H and δ18O values range from − 34.9‰ (LK16) to − 1.9‰ (LK3) and − 5.30‰ (LK16) to + 0.70‰ (LK3), respectively (n = 40) (Table S2). Lake waters plot to the left of the Global Meteoric Water Line (GMWL, red solid line on Fig. 4) on a regression line described by δ2H = 8 δ18O + 1035, and the Cape Grim local meteoric water line (LMWL, δ2H = 6.8 δ18O + 6.6536 shown as a blue dashed line in Fig. 4). The average δ2H and δ18O values for the 40 lake waters is − 20.1‰ and − 2.83‰, respectively, which indicates they are slightly more enriched in 18O than the amount weighted rainfall values for Cape Grim (− 20.3‰ and − 3.97‰). The δ18O and δ2H values are not correlated with Cl (⍴ = 0.50 and 0.47, respectively) and do not plot on a mixing line towards seawater (shown as a yellow diamond in Fig. 4).
Difference between Macquarie Island Lake water samples (δ2H = 5.6 δ18O − 4.3), the global meteoric water line (GMWL, δ2H = 8 δ18O + 10) and Cape Grim local meteoric water line (LMWL, δ2H = 6.8 δ18O + 6.65). (Cluster groupings from Fig. 3a).
87Sr/86Sr ratios range from 0.70694 (LK29) to 0.70908 (LK38) with an average value of 0.708341 (n = 10). 87Sr/86Sr ratios were mildly positively correlated with Sr concentrations (⍴ = 0.64, p = 0.04) suggesting a possible source of Sr with high 87Sr/86Sr ratios. δ13CDIC values for the lake waters are highly variable and range from − 23.1‰ (LK37) to − 1.5‰ (LK29) with an average of − 11.9‰ (n = 40). The DOC concentrations range from 0.7 (LK14) to 6.8 (LK44) mg l−1 with an average of 2.4 mg l−1 (n = 40). The δ13CDOC values are more consistent and range from − 23.7‰ (LK35) to − 36.7‰ (LK19) with an average of − 28.2‰ (n = 40).
Spatial variation in lake water chemistry
The ionic concentration of the lake waters are controlled spatially, where Cl concentrations are highest along the west coast, and lowest in the centre, north and east coast of the island (Figs. S1 and 5a). Cl concentrations are not related to elevation (⍴ = 0.09, p = 0.6) or lake area (⍴ = 0.03, p = 0.8). The spatial distribution of K, SO4, Br, Na and Sr concentrations follows a similar pattern (Figs. S1–S3, Fig. 5b–d). Sr is significantly positively correlated with Cl (⍴ = 2.2 × 10–7), Br (⍴ = 1.5 × 10–5), Mg (⍴ = 7.2 × 10–5) and Na (⍴ = 2.6 × 10–6, Table S3). Overall, there is a general trend of decreasing Cl, Na, SO4 and Sr with increasing distance from the west coast, whilst elements such as Fe, Ca, F and Si and isotopes such as δ18O, δ13CDIC and δ13CDOC do not show a clear trend (Fig. 5e,h,i,k,l).
Distribution of ion concentrations and isotopes relative to distance from the west coast. Note the generally high concentrations of (a) Cl, (b) Na, (c) SO4 and (d) Sr in the lakes from Cluster 1 (green) are located close to the west coast. Concentrations of (e) Fe are mixed in lake waters impacted by sea-spray, whilst concentrations of (f) Ca, (g) F and (h) SiO2 are low in these lakes and increase with distance from the west coast, particularly in samples identified as having undergone high water–rock interaction. Lakes in Cluster 1 located close to the west coast contain higher (i) δ18O and (j) 87Sr/86Sr ratios and relatively low (k) δ13CDIC values. δ13CDOC values (l) show no significant relationship with distance from the west coast (p > 0.05).
Sixteen of the lakes have SiO2 values of 0.001 mmol l−1, whilst LK16 contained the highest concentration (0.087 mmol l−1). LK16 is located at the second lowest elevation (at 95 m a.s.l) of lakes in this study (Table S2). Ca concentrations range from 0.02 mmol l−1 (LK8) to 0.29 mmol l−1 (LK9). LK8 is located at the third lowest elevation of lakes sampled in the study (Table S2). Three lakes have F concentrations of 0.0003 mmol l−1, whilst the highest F concentration is identified in LK16 at 0.0058 mmol l−1. SiO2, Ca and F are not significantly related to distance from the west coast (⍴ = 0.10, 0.22 and 0.08, p = 0.52, 0.18 and 0.61 respectively) (Fig. 5h,f,g) or Cl concentration (⍴ = 0.16, − 0.14 and 0.23, p = 0.33, 0.37 and 0.15 respectively). Fe and F are, however, significantly negatively correlated with elevation (⍴ = − 0.49, p = 1.5 × 10–3 and ⍴ = -0.41, p = 9.4 × 10–3 respectively). Ca and F are positively correlated with SiO2 (⍴ = 0.78 and 0.49, p = 2.4 × 10–9 and 1.3 × 10–3 respectively, Table S3) suggesting that their source is different from that of SO4, Cl, Mg and K (see also Fig. 2).
Source: Ecology - nature.com
