Study Area & Methods

Study Area

      Study sites were located between kilometer 14 and kilometer 15 along the ice road, 15 km's outside the town of Burton, B.C. Annual total precipitation ranged from 900 to 1000 mm (25-50% falls as snow), increasing with elevation. Annual air temperature ranged from 2 to 4 ºC. All study sites had the same soils: Dystric Brunisol. Average slope was 30% across the sites with micro-topograhical attributes randomly dispersed.          
      In September of 1996, target basal areas were selected at 25% and 50% of the original stand (36 m2/ha). Trees were hand felled and skidded using a 1987 Komatsu 37E crawler tractor. In April of 1998, 1 year-old, container grown (1-0 PSB 313A) seedlings were planted.




Methods

     Three retention levels (25%, 50% and 100%(control)) as a percentage from the original basal area (36 m2/ha) were each randomly assigned to one hectare treatment blocks (100m x 100m). The control was replicated three times, while treatments were replicated twice. 
     Sixteen subplots (four rows of four) were assigned in each one hectare treatment block. Subplots were systematically placed in the middle each block and were 15 m apart (leaving a 20 meter buffer on the edges). The middle two rows (2 rows of four) were selected to reduce edge effect. One tree per species was selected to be measured based on dominance and effect of residual trees within 4 m of the subplot center (Fig. 2). Measured tree characteristics are summarized below (Table 1).


Light Measurements

     Light measurements were taken at each sample tree with a LAI 2000 Plant Canopy Analyzer. Two measurements were taken (East and West) with 180 degree restrictors to avoid error with direct sunlight. Values were compared to an open unit approximately 1 kilometer away and averaged to get a full 360 degree indication of the light level and competition.  Diffuse Non-interceptance Light (DIFN) was the estimator of light levels and Leaf Area Index (LAI) was the estimator of competition.



Volume Equations

Volume equations were formulated using increment data from tree's growing in the vasincity of the study area (less than 15 km's). Total height (cm) and diameters at 25 cm intervals for trees less than 2 meters and 50 cm intervals for trees greater than 2 meters were taken for each species. Data was fitted to horner's equations using Proc ALIN in SAS. Horner's equation has the following form: 
                                              


                                              [1]     Vol = (D^a)/b-(c/Ht)

Where Vol is the indivual tree volume (in cm3), D is diameter (cm) at breast height when trees are greater than 2 m and collar diameter (cm) when trees are less than 2 meters. Ht is the height of the tree in centimeters, a,b,c are fitted parameters from data set (Table 2). Growth reponse was determined by the percent increase in volume to reduce co-variation of tree size.





Statistical Methods

(1i) Individual species growth (continuous data)  was compared across the three levels of retention (categorical) using an ANOVA to determine if retention level had an effect on growth. A data table (classed by species) was created with retention level (categorical data) as the predictor and growth (continuous) as the response. Tukey's pair wise comparisons determined which rentention levels differed in growth. 

(1ii) The response of growth between species within each rentention level was found using an ANOVA. Three data tables were subseted for each retention and classed by species. The means of each species were compared with each other. Tukey's pair wise comparisons were used to determine which species differed. Individual T-test's were used to determine how much of a difference was found.

(2i) Light (%Full sunlight) was compared across each retention level using a data table that used retention level as the predictor variable and light as the response variable. An ANOVA was used to determine the if a difference existed. Light was categorized in 10% increments and ploted against the percent found in each retention level.


(2ii) Light environments were stratified into four distinct light classes (width of 20% (categorical)) based on a percentage of full sun (DIFN). A data table consisting of light class(classed by species) as the predictor variable and percent increase in volume as the reponse variable was constructed. In each light environment the three species growth responses were compared with an ANOVA and subsequently determined which species were different with Tukey's pair wise comparisons. 


(2iii) The reponse of species growth (% increase in volume) was compared to the leaf area index (contiunous variable), taken at each sample tree. Three logorithmic equations were made and thresholds were determined based on intersection of curves. Equations were made based on biologcial understanding (logorithmic) and fitted using PROC NLIN in SAS. Each equation had the following form:

                   [2]             Percent Increase in Volume =  βo + β1*ln(LAI)

Significance was determined at the 0.05 alpha level. All graphs were computed in R. All statistical analyses were computed in SAS.


Disclaimer: this is a class exercise based on modified or randomly generated datasets