Results/Discussion

Both CDA through geologic unit and CDA through geomorphic class explained ~75-80% of the variation between classes within the first two discriminant functions. In both classification methods, lake groupings varied in terms of overall nutrient concentration (TN and DOC relatively together) and overall cation concentration (Mg, Ca, K relatively together). When analyzing by geomorphic class, one grouping (RLL) contained much less sodium than other lake classes; manual inspection of the dataset showed that this was consistent across the sample set, rather than the effect of an outlier. Both classification methods separated out highly productive (high-nutrient) lake classes (respectively: Gp and O in surficial units. and GEO-PPP and OTB in geomorphic units).

Note: all chemical parameters listed in Data Walkthrough were used, however only the top ten most contributing are shown in each of these two plots

When the extended suite of chemistry (Figures 13 and 14) was used in the analysis, sulfur compounds became one of the top ten distinguishing chemical features between classes, regardless of whether geomorphic class or geologic unit was used. Lakes existing on heavily glacially-influenced terrain (moraine lakes: Mp/Mh/Mk geological units, ML geomorphic class) had higher concentrations of sulfur compounds (SO) than any other lake grouping. This is likely due to the mineralogical composition of the sediments immediately surrounding these lakes, and potentially the flow paths which water has taken before arriving at a lake; water bodies with higher groundwater contributions relative to rainwater, and groundwater which has travelled a longer distance in contact with sediment ("deeper"/"older" water) are likely to have higher sulfur compound concentrations.

Further supporting this, chloride concentration (Cl) is located nearly parallel to, and in the opposing direction to, sulfur concentration in the geological class CDA. Chloride is the most easily dissolved anion in terms of groundwater flow paths; groundwater which has not travelled in contact with sediment for long distances ("shallower"/"younger" water) is relatively enriched in chloride relative to sulfur. Surficial geology has previously been documented to affect groundwater-related mineral constituents of lakewater (Larsen et al., 2017).

While both geological and geomorphic classifications successfully differentiated lake chemistry, each method may be more suited to specific research questions. In terms of lake productivity, and therefore carbon flux estimation, identification of glaciofluvial and lacustrine floodplain sediments (geological classification), and the highly productive polygonal patterned ponds which form on them (geomorphic classification) with little discrimination required between remaining lake types due to relatively similar nutrient profiles, may be achieved with both methods. (Polygonal patterned ponds, which are very small, shallow (meaning warm, with high light availability), and young in terms of formation age (forming from melt of ice in surface sediments which contain organic matter) (see Kokelj et al., 2023), are known to be highly productive systems.)

Classification of lakes by geological unit discriminates more finely in terms of inorganic chemical concentrations (cations and anions, visible from the relative angle and presence of the top ten factors in Figure 13). Geomorphic classification discriminates between lakes in terms of nutrient concentrations, and the specific ratios between them (i.e. the inclusion of NO3 within the top ten, and relative separation of DOC and TN in Figure 14). Lake colour (ColourAp, ColourTrue), which is a metrics of coloured organic compounds (i.e. tannins), and correlates to the amount and type of organic matter input into these systems, was also a stronger differentiating factor in geomorphically classified lakes than geologically classified lakes.

Both geological and geomorphological classification methods successfully differentiated lake chemistry, and differentiated the high-nutrient, high-productivity lake types which are expected to produce the most methane. (These being polygonal patterned ponds geomorphically, and glaciofluvial floodplain and lacustrine plain sediments geologically.) The use of these methods of classification would allow for lakes to be inventoried by class at landscape to regional scales. This in turn would allow for models of lake productivity and lake methane fluxes to be more accurately generalized at these scales, given they can be modified to the chemical composition of specific lake classes, which each have known abundances. Outside of this, the two classification methods highlighted different aspects of lake variation, and may have different alternative applications.

For questions of water potability for communities, geological classification may be useful, as it is more sensitive to variation in cation and anion concentrations.  For questions of function within the umbrella of productivity, such as the specifics of nutrient limitation and utilization, geomorphic classification is preferrable, since it distinguishes between nutrients more strongly. This suggests that inorganic chemistry may be more strongly influenced by the sediments surrounding a given lake, whereas ecological behaviour and nutrient cycling, may be driven by factors more localized and captured by morphological assessment, such as basin depth and morphology.