RESULTS

Date of ice-out increased with increasing lake elevation on the east slope (r² = 0.259, N = 16, P = 0.044) and on the west slope (r² = 0.621, N = 39, P < 0.001) of the park. East-slope subalpine lakes iced-out earlier that did west-slope subalpine lakes (P = 0.01). This result occurred even though the average elevation of east-slope subalpine lakes was higher than the average elevation of west-slope subalpine lakes (Table 2.1).

On average, NOCA lakes were relatively cold, neutral in pH, and low in dissolved substances and concentrations of nitrogen and phosphorus (Table 2.5). However, considerable variation was observed for most variables. Water temperature, pH, alkalinity, conductivity, total Kjeldahl-N, and ammonia-N of west-slope lakes decreased in concentration and nitrate-N increased in concentration with increased elevation (Table 2.6). Only total phosphorus and orthophosphate-P did not change in concentration relative to elevation (Table 2.6). However, when three lakes (two alpine and one subalpine) receiving turbid-glacial outwash were deleted from the data set, total phosphorus increased with decreasing elevation (Table 2.6). Changes in vegetation zone were related to elevation as well. East-slope subalpine lakes were significantly higher in pH, alkalinity, total phosphorus, and total Kjeldahl-N than were west-slope subalpine lakes (Table 2.7).

Table 2.5. Average water quality of west-slope and east-slope NOCA study lakes.

*detection limits

Table 2.6. Regression analyses of water quality variables and begetation zone as independent variables and elevation as the dependent variable for west-slope NOCA lakes.

Table 2.7. Comparisons of water qualities of west-slope subalping lakes (n=16) and east-slope subaline (n=9) lakes, NOCA.

*Mann-Whitney U-test

Based on the second discriminant analysis, using six water quality variables and elevation as independent variables with vegetation zone as the dependent variable, 98.2% of the among-group variation was found to be associated with the first two canonical variates; 93.3% was associated with variate 1 and 4.9% with variate 2. Canonical variate 1 was positively correlated with elevation and negatively correlated with alkalinity, conductivity, total Kjeldahl- N, and pH (Table 2.8). Canonical variate 2 was most strongly and positively correlated with orthophosphate-P and nitrate-N (Table 2.8). The effectiveness of this model in placing lakes into the correct vegetation zone-based classification categories was somewhat problematic, however, since 13 of the 58 lakes (22.4%) were misclassified.

Table 2.8. Correlation of independent variables and canonical variates one and two determined during the second discriment analysis.

Collectively, shallow lakes (<10 m in maximum depth) exhibited a wide range of values for alkalinity and conductivity, and concentrations of total Kjeldahl-N and total phosphorus. A greater percentage of shallow lakes maintained high values of the variables than did deep lakes (Table 2.9).

Table 2.9. Number of shallow (<10 m) and deep (>10 m) lakes exhibiting elevated conductivity, alkalinity, and concentrations of total Kjeldahl-N and total phosphorus. Possible numbers of lakes: shallow (38), deep (19).

Variable	Concentration or Level	Number of Percentage of Lakes
		<10 m	%	10 m	%
Conductivity	25 µS/cm	14	37	3	16
Alkalinity	3 mg/L	12	32	3	16
Total Kjeldahl-N	0.08 mg/L	10	26	1	5
Total Phosphorus	0.008 mg/L	14	37	3	16

Geology did not play a major role in segregating most NOCA lakes based on water quality. For example, water quality of lakes in granitic or gneiss basins, the primary geologic substrates, did not differ (see Lomnicky 1996). The few low-forest lakes in watersheds with greenstone outcroppings were higher (P = 0.025) in alkalinity and conductivity than were low- forest lakes in either gneiss or granite bedrock watersheds (Table 2.10).

Table 2.10. Comparisons of conductivity and alkalinity of low-forest west-side lakes in basins with gneiss or granite and those in greenstone (n=3).