Devlin and Finkelstein, Recent ecological changes in aquatic ecosystems in two distinct physiographic regions of Sirmilik National Park, Nunavut, Canada

Electronic supplementary material

Methods used for collection and preparation of water samples for chemical analysis at National Lab for Environmental Testing (NLET, Burlington, Ontario, Canada):

Grab samples (2 x 1000-mL) were taken from helicopter floats after bottle conditioning and stored at 4°C in a dark cooler. Samples for filtered analyses were prepared in the field within 48 hrs of sampling. Sample water was filtered through 0.45-m cellulose-acetate membrane filters for dissolved organic carbon (DOC), dissolved inorganic carbon (DIC), total dissolved nitrogen (= Total Nitrogen Filtered, TNF) and total dissolved phosphorus (= Total Phosphorus Filtered, TPF). For total particulate phosphorus (TP-PART) samples, 0.45-m cellulose-acetate membrane filters which had filtered 250 mL (minimum) of sample water were transferred to separate glass bottles each containing 1.0 mL H2SO4 (30%) preservative. For unfiltered total phosphorus (TP-U), 50 mL of unfiltered sample water was added to glass bottles containing 0.5 mL of H2SO4 (30%) preservative.

For chlorophyll-a sample collection, a minimum of 750 mL of sample water was filtered through Whatman 4.7-cm diameter GF/C glass-fibre filters, and 0.2 mL of magnesium carbonate suspension (1% w/v) was pipetted ontothe filter paper. Filter papers were wrapped in aluminum foil and stored in the freezer until shipping. Samples were taken between July 26-31, 2008, and were submitted to NLET on August 5, 2008.

Supplementary Table 1Summary of variable significance testing and Canonical Correspondence Analysis (CCA) results. CCAs included diatom assemblages from 20 lakes and ponds in SNP (see Table 1 in the main text for locations). Inter-set correlations are correlations of environmental variables with the species-weighted-average site scores(ter Braak 1986)

Single-variable CCAs / CCA using 8 canonical variables
Real data / Monte Carlo test, 998 runs
(randomized data) / Inter-set correlations
Spp-Envt Corr. / Mean / Minimum / Maximum / P / Axis 1 / Axis 2 / Axis 3
Depth / 0.913 / 0.694 / 0.517 / 0.874 / 0.001 / 0.827 / -0.033 / 0.253
Distance From Coast / 0.924 / 0.696 / 0.495 / 0.909 / 0.001 / -0.838 / -0.299 / 0.170
Elevation / 0.922 / 0.692 / 0.499 / 0.868 / 0.001 / 0.875 / 0.053 / 0.007
pH / 0.913 / 0.693 / 0.498 / 0.881 / 0.001 / -0.890 / 0.107 / -0.065
Specific Conductivity / 0.962 / 0.692 / 0.483 / 0.917 / 0.001 / -0.921 / 0.242 / -0.133
Surface Area / 0.949 / 0.689 / 0.511 / 0.909 / 0.001 / 0.843 / 0.360 / 0.007
Turbidity / 0.940 / 0.704 / 0.370 / 0.953 / 0.002 / -0.736 / 0.498 / 0.269
Water Temperature / 0.873 / 0.691 / 0.525 / 0.870 / 0.001 / -0.791 / -0.048 / 0.031

Supplementary Table 3 Summary statistics for CCA of surface diatom assemblages at 20 sites in SNP. Environmental variables used in the analysis are listed above. Eigenvalues () indicate the amount of variance extracted by each axis. Total variance (inertia) in the species data = 4.1485

Axis 1 / Axis 2 / Axis 3
Eigenvalue () / 0.819 / 0.687 / 0.407
Variance in species data (% of variance explained) / 19.7 / 16.6 / 9.8
Cumulative % explained / 19.7 / 36.3 / 46.1
Pearson Correlation, Spp-Envt* / 0.991 / 0.986 / 0.907

*Correlation between sample scores for an axis derived from the species data and the sample scores that are linear combinations of theenvironmental variables