Abiotic and biotic variables of lakes
There were significant differences in average total length and weight among lake trout collected throughout 2017 annual sampling from L260, L223, L224, and L373 (ANOVA, p ≤ 0.05) (Table 1b; Figure S4). Average total length and weight of lake trout in Lake 260 were significantly greater than average total length and weight of lake trout in L223, L224, and L373 (Tukey HSD, adj-p ≤ 0.0001 for all comparisons). Furthermore, average total length and weight of lake trout in Lake 223 were greater than average total length and weight of lake trout in L224 (Tukey HSD, adj-p ≤ 0.002 for both comparisons). There were no significant differences in average total length and weight among L224 and L373 and among L223 and L373 (Tukey HSD, adj-p> 0.05 for all comparisons). Although there were differences in length and weight, average condition factors did not significantly differ among populations (ANOVA, p = 0.1) (Table 1b; Figure S4).
Total zooplankton abundance ranked from highest to lowest was as follows: L223 > L260 > L224 > L373 (Table 1c). L223 and L260 also had the highest zooplankton species richness (38 species), whereas species richness was lower in L224 (29 species) and L373 (26 species) (Table 1c). The most abundant zooplankton within all lakes were rotifers (e.g. Keratella cochlearis ,Kellicottia bostoniensis , Polyarthra remata ,Polyarthra vulgaris , Kellicottia longspina ) and copepods (e.g. cyclopoida nauplii ) (Figure S5). L373 had the highest H’ index of 2.50, followed by L260 with 2.08, then L223 with 1.93, and finally L224 with 1.67 (Table 1c). L373 and L260 also had the greatest evenness of zooplankton species with equitability index values of 0.77, followed by L223 and L224 with equitability values of 0.53 and 0.50, respectively. Thus, L373 and L260 contain the most diverse, evenly-distributed zooplankton communities. Abundance of all species across all lakes can be found in Additional File 1: Table S7.
Average water quality parameters are presented in Table 1d and visualized in Figure S6. There were no significant differences in average suspended phosphorus, suspended nitrogen, suspended carbon, and total dissolved phosphorus among lakes (Tukey HSD, adj-p> 0.05 for all comparisons). However, L223 and L260 had significantly higher levels of total dissolved nitrogen than L224 and L373 (Tukey HSD, adj-p ≤ 0.0001 for both comparisons). Chlorophyll-a levels were significantly lower in L224 compared to Lakes 223 and 260 (Tukey HSD, adj-p > 0.03 for both comparisons), while chlorophyll-a levels did not significantly differ among L223, L260, and L373 (Tukey HSD, adj-p > 0.05 for all comparisons). However, there was a visible trend of increased chlorophyll-a in L223 and L260 (Figure S6). Average dissolved oxygen levels in L260 were significantly lower compared to all other lakes (Tukey HSD, adj-p ≤ 0.0001 for all comparisons). Average alkalinity was significantly higher in L373 compared to all other lakes (Tukey HSD, adj-p ≤ 0.0001 for all comparisons), whereas alkalinity was significantly lower in L224 compared to all other lakes (Tukey HSD, adj-p ≤ 0.0001 for all comparisons). Furthermore, average pH was significantly higher in L373 as compared to L224 and L260 (Tukey HSD, adj-p ≤ 0.01 for all comparisons ) but pH did not significantly differ among L223, L224 and L260 (Tukey HSD, adj-p> 0.05 for all comparisons). Average temperature was comparable among the lakes; however, water temperature of L373 was significantly lower than in L223 and L224 (Tukey HSD, adj-p ≤ 0.03 for all comparisons) whereas temperature did not significantly differ among L223, L224, and L260 (Tukey HSD, adj-p> 0.05 for all comparisons). Finally, average conductivity was significantly lower in L224 compared to all other lakes (Tukey HSD, adj-p ≤ 0.02 for all comparisons).
PCA of 11 variables for the four lakes was used to identify variation in water quality characteristics. Proportions of variance explained by PC1, PC2, and PC3 were 50.39%, 29.53%, and 20.08%, respectively. PC1 and PC2 explained the maximum amount of total variance (79.92%). Variable loadings on each PC are presented in Additional File 1: Table S8 and are graphically represented in a PCA biplot (Figure 6). PC1 had strong positive loadings from suspended nitrogen (0.42), suspended carbon (0.38), conductivity (0.38), and chlorophyll a (0.36) (Figure 2; Additional File 1: Table S8). The highest scoring lakes for PC1 were L223 (1.18) and L260 (1.68), while the PC1 score was weakest for L373 (0.61) and was strongly negative for L224 (-3.47). Furthermore, the strong negative PC1 score of L224 and strong negative PC1 loading of total dissolved phosphorus (-0.39) suggest that L224 had high total dissolved phosphorus relative to lakes scoring positively on PC1 and that total dissolved phosphorus likely drove the variation of L224 from other lakes.
PC2 had strong positive loadings from pH (0.55), alkalinity (0.42), and dissolved oxygen (0.34) as well as negative loading from chlorophyll-a (-0.24) (Figure 6; Additional File 1: Table S8). L223 was the only lake with a positive PC2 score (2.49) out of the lakes of interest (Figure 6). Overall, PCA suggested that L260 and L223 were most similar in terms of water quality characteristics while L224 and L373 varied from all other lakes (Figure 6). Hierarchical clustering analysis confirmed that L223 and L260 were most closely related based on water quality characteristics whereas L224 and L373 clustered apart from L223 and L260 (Figure S7).