METHODS

Fifty-eight lakes were sampled from 1989 through 1993 while they were free of snow and ice (Table 2.2). Some lakes were sampled one to three times, whereas others were sampled up to 12 times during the 5-year study. The lakes were located in remote mountainous terrain, and were accessed by hiking or by helicopter. Sampling was conducted from an inflatable boat over the deepest spot in each lake. Water temperatures were determined at a depth of 1 m using an Omega HH70 hand-held thermometer. Water samples were collected 1 m below the surface of each lake using a 1.5-L Van Dorn-style bottle. Filtered (0.7-µm prewashed Watman GF/C filters) samples were placed in 1-L acid-washed, high-density polyethylene bottles for nutrient analyses. Unfiltered samples collected in 250 ml acid-washed polypropylene bottles were used for analyses of pH, alkalinity, and conductivity. To the degree possible, water samples were taken just before leaving the field to minimize sample deterioration. If samples were kept overnight in the field, the bottles were placed out of the light during the day or in dark bags and kept in the coldest spot that could be found (e.g., inlet stream, snow bank). When departing the field, the samples were wrapped with insulation to maintain coolness, transported out of the field, frozen, and then shipped to the Cooperative Chemical Analytical Laboratory, Oregon State University. Chemical analyses included pH, alkalinity, conductivity, total Kjeldahl-N, nitrate-N, ammonia-N, total phosphorus, and orthophosphate-P (Table 2.3).

Table 2.2. Elevations, surface areas, maximum depths, and vegetation zones and the number of times each lake was sampled (1989-93) for NOCA study lakes east and west of the hydrologic divide.

Alpine (1), subalpine (2), high-forest (3), and low forest (4)

Table 2.3. Analytical procedures used by the Cooperative Chemistry Analytical Laboratory, Oregon State University.

Dates of ice-out, defined as the time when the lake surfaces were virtually ice-free, were estimated for most lakes from aerial surveys conducted in 1989 when the field crews were helicoptered to and from sampling locations. The estimated dates of ice-out likely were within 1 to 2 weeks of the actual dates in some cases. For a few lakes not covered by the aerial surveys, estimated dates of ice-out were obtained from backcountry park rangers.

Average values for the water-quality variables were calculated using the STATS module of SYSTAT (Wilkinson 1987). Average pH was calculated from hydrogen ion activity. Regression analyses were used to evaluate dates of ice-out and changes in water quality with increased lake elevation for lakes west of the hydrologic divide using the MGLH module of SYSTAT. West-slope NOCA lakes were used because samples were available from a greater range of lake elevation (i.e., alpine to low-forest lakes), than were east-slope lakes (primarily subalpine lakes). Statistical comparisons of dates of ice-out and the water qualities of east-slope and west-slope lakes in the subalpine vegetation zone were conducted using the Mann-Whitney test (NPAR module of SYSTAT). Level of statistical significance was = 0.05. Discriminant analysis (NCSS vers. 5.03, Hintze 1992) was used to examine how the water quality variables as well as elevation, lake depth, and lake surface area (independent variables) would predict the placement of lakes into four classification categories (dependent variable) based on the four vegetation zones. This was elucidated during the analysis by: 1) the construction of a classification matrix according to the assignment of lakes to the classification categories; 2) determination of classification accuracy by calculation of the percent reduction in classification error due to the independent variables; and 3) determination of the number of lakes misclassified. The discriminant analysis also calculated three canonical variate scores for each lake, and calculated the percent variation attributable to each variate, as well as the correlations between each variate and each independent variable. An initial analysis was performed using all independent variables (Table 2.4). A second analysis was performed using only independent variables found to have significant influence during the first analysis.

Table 2.4. Identification of independent variables having significant influence during the initial discriminant analysis using all eleven variables.