Collection Mémoires et thèses électroniques
AccueilÀ proposNous joindre

4 Materials and methods

Table des matières

This study was conducted in the black spruce-feathermoss ecological domain of eastern Quebec, Canada (49,94° N to 51,30N; 66,77°W to 69,56°W),north of Baie-Comeau near René Levasseur Island and the hydroelectric reservoir known as Manic 5 in the Manicouagan Côte-Nord region (Fig. 5). The study area is found in the Manic reservoir landscape unit of the ecological district ‘Hautes collines du réservoir Outardes’ (6i; Robitaille and Saucier 1998), and is characterized by moderate hilly topography, high and round hills with moderate and regular slopes. Mean altitude is 499 m. The bedrock consists of sedimentary rock of crystalline nature and metamorphic gneiss; till and fluvioglacial deposits are common (Robitaille and Saucier 1998). The region has a subpolar subhumid continental climate with a mean annual temperature of -2.5 to 0.0°C. Mean annual precipitation is 1300 mm with 350 to 400 cm of snow. The growing season is only 140days.

The main plot was divided into 20 sub-plots of 4m2, arranged in a cross (Fig. 6). The second sub-plot of each branch of the cross was used to evaluate the percent cover of shrubs, herbs, mosses, and lichens (macrolichens; less common crustose lichens, fungi, and liverworth were not included). We then calculated a mean for the site from the four subplots. The pre-established cover classes used were: 0-5 %, 5-10 %, 10-20 %, 20-40 %, 40-60 % 60-80 %, 80-100 % and 100% (area entirely covered by the species). Tree seedlings (>5cm and <1.3 m of height) were counted by species in a 1 m2 plot in the north-west corner of each 4 m2 subplot.

Four stand cover types were identified: 1)Black spruce stands (SbSb) arecomposed mainly of black spruce, with cover >75% of the conifer basal area; 2)Black spruce-balsam fir stands (SbBf) where black spruce cover is 50-75% of conifer basal area and balsam fir is the second dominant species; 3) Balsam fir-black spruce stands (BfSb) where balsam fir cover is 50-75% of conifer basal area and black spruce is the second dominant species; and 4) Balsam fir stands (BfBf) are mainly composed of balsam fir with cover >75% of the conifer basal area.Age classes were defined as follows: 50 yrs (41-60 years),70 yrs (61-80 years),90 yrs (81-100 years), and 120 yrs (101 years and more).

The sampling sitewas composed of a circular main plot of 400 m2 (radius of 11.28 m) divided into 20 sub-plots of 4 m2. The second sub-plot (in gray) of each branch of the cross was used to evaluate ground cover of vascular and nonvascular plants (%). Tree seedlings were counted by species in a 1 m2 plot in the north-west corner of each 4 m2 subplot.

A series of descriptive statistics, SAS software version 9.1 (SASInstitute 2005), were used to evaluate site characteristics according to environmental variables, ericad cover, and soil characteristics.Mean values of structural parameters mean values were evaluated for all stands without stand cover type discrimination. The environmental variables were: canopy openness, humus thickness, slope, drainage, mineral soil depth, soil texture (sand, silt, loamy sand, sandy loam, and silt loam), black spruce merchantable tree diameter at breast height, and dominant black spruce mean age.The ericads were:Rhododendron groenlandicum(Oder) Kron. and Judd.,Vaccinium angustifoliumAit.,Vaccinium vitis-ideaL.,Vaccinium myrtilloidesMitchx.,andKalmia angustifoliaL.The soil variables were:humus pH, humus carbon: nitrogen ratio,humus carbontotal concentration, humus nitrogentotal concentration, humus phosphorus totalconcentration, humus potassium totalconcentration,mineral soil pH, mineral soil carbon: nitrogen ratio, mineral soil carbon total concentration, mineral soil nitrogen total concentration, mineral soil potassium total concentration, mineral soil calcium total concentration, mineral soil magnesium total concentration, mineral soil sodium total concentration, mineral soil acidity (H+), mineral soil cation exchange capacity, mineral soilbase saturation,sand, silt, and clay.

Analysis of variance (SAS Institute 2005) was applied to evaluate stand cover type characteristics and to quantify the presence of ericaceous species (Rhododendron,Vaccinium, andKalmia) in stands originating from natural disturbance in the North Shore forests (SbSb, SbBf, BfSb, and BfBf).Means were compared by Tukey’s test according to stand cover types. Principal component analysis (PCA), XLSTAT-Pro software (Addinsoft 2006), was performed to visualize stand cover types in relation to environmental and soil characteristics.PCA was also an aid to reduce variable redundancy. Principal component analysis used the correlation among the environmental variables to develop small sets of components that empirically summarize the correlations among those variables. Variability was measured using a correlation matrix. Varimax, an orthogonal rotation, was used after extraction to maximize high correlations and minimize low ones (Tabachnick 2001).

A series of descriptive statistics and a series of multiple regression analyses (SAS institute 2005) were used to study the relationship among site characteristics and the presence of ericaceous shrubs.The stepwise procedure was used as an aid for selecting and assessing models for ericaceous species. This statistical regression was used to find the best predictive equation for the ericaceous cover. Several variables (environmental, plant and soil variables) were combined into a predicted variable to produce the highest correlation with the single variable, the ericaceous species (in a method described by Tabacknick 2001).The plant variables were:Rhododendron groenlandicum(Oder) Kron. and Judd.,Vaccinium angustifoliumAit.,Vaccinium vitis-ideaL., andVaccinium myrtilloidesMitchx.,Kalmia angustifoliaL.,Gaultheria hispidulaL. Muhl.,Rubus chamaemorusL., Cornus canadensis L.,Clintonia borealis(Ait.) Raf.,Linnaea borealisL.,Maianthemum canadenseDesf.,Oxalis montanaRaf.,Coptis groenlandica(Oeder)Fern.,Lycopodiumspp.,Trientalis borealisRaf.,Carexspp.,Equisetumspp.,Pleurozium schreberi(Brid.) Mitt.,Ptilium crista-castrensis(Hedw.) De Not.,Sphagnumspp.,Hylocomium splendens(Hedw.) B.S.G.,Dicranumspp.,Polytrichumspp.,Cladina stellaris(Opiz) Brodo,Cladina rangiferina(L.) Nyl.,Cladina mitis(Sandst.) Hustich.Qualitative variables (stand cover types and soil textures) were converted into a set of dichotomous variables by dummy variables coding with 1s and 0s (Tabachnick 2001). Variables were selected for their correlation with the dependant variable (ericaceous species) with an error risk of α = 0.20. Variables were then included or omitted from the equation according to a probability level of 0.05. The relationship among environmental variables, included in the model, and the presence of ericaceous shrubs were evaluated using a series of descriptive statistics. Collinearity was checked with an option for model checking within PROC REG (SAS institute 2005). Ericaceous species were arcsine transformed, “angular transformation”, prior to analysis (Zar 1974). Slope expositions were cosine transformed (Zar 1974). Principal component analysis, XLSTAT-Pro software (Addinsoft 2006), was used to identify ericaceous shrub’s distribution according to environmental and soil characteristics. Canonical correspondence analysis (CCA), PC-ORD software (McCune and Mefford 1999), was another statistical method used to underline differences among environmental and soil nutrient variables and the presence of ericaceous shrubs. This method seeks the combinations of environmental variables or soil variables that best explain the variation of the ericaceous cover (Tabachnick 2001).

Analysis of variance (SAS Institute 2005) was used to evaluate regeneration and tree growth measures for each stand cover type (SbSb, SbBf, BfSb, and BfBf). Principal component analysis, XLSTAT-Pro software (Addinsoft 2006), was used to visualize individual ericad species distribution according to regeneration and growth parameters. Canonical correlation analysis was also used to underline differences among individual ericad species, as well as regeneration of black spruce and balsam fir, andgrowth parameters of black spruce.

An analysis of variance (SAS Institute 2005), with Tukey’s test, was applied to the mean percent cover ofRhododendron, Vaccinium,andKalmiawithin stand cover types, within ecological site types defined by the MRNQ typology and within ecological site types defined in typology of Côté (2006).Three typology classifications were used. One was based on stand cover types (SbSb, SbBf, BfSb, and BfBF). The second was the typology formulated by the MRNQ. MRNQecological types are as following: RE10, black spruce-lichen type, shallow soil, variable soil texture, xeric to hydric drainage; RE11, black spruce-lichen type, shallow to deep soil, coarse soil texture, xeric-mesic drainage; RE12, black spruce-lichen type, thin to deep soil, medium soil texture, mesic drainage; RE20, black spruce-feathermoss or -ericad type, shallow soil, variable soil texture, xeric to hydric drainage; RE21, black spruce with feathermoss or ericad species, shallow to deep soil, coarse soil texture, xeric-mesic drainage; RE22, black spruce stand with feathermoss or ericad species, shallow to deep soil, medium soil texture, mesic drainage; RE25, black spruce stand with feathermoss or ericad species, shallow to deep soil, medium texture, subhydric drainage; RS20, balsam fir/black spruce type, shallow soil, variable soil texture, xeric to hydric drainage; RS22, balsam fir/black spruce type, shallow to deep soil, medium soil texture, mesic drainage; RS25, balsam fir/black spruce type, shallow to deep soil, medium soil texture, subhydric drainage; RS2A, balsam fir/black spruce stand, with seepage.The third was the typology recently formulated by Côté (2006). Côté (2006)ecological types are as following: Type 1, young even-aged black spruce stands; Type 2, even-aged black spruce stands; Type 3, uneven-aged black spruce-balsam fir stands; Type 4, uneven-aged balsam fir-black spruce stands; Type 5, open black spruce stands; and Type 6, non-regenerated black spruce/balsam fir stands.An ANOVA (Tukey’s test) was also carried out to evaluate the mean percent cover ofRhododendron, Vaccinium,andKalmiawithin stand age classes.

Cluster analysis,using Ward’s method,and principal component analysis (XLSTAT 2006) were used to identify the most important vegetation associations found in the forests of the North Shore region. Cluster analysis allowed detection of vegetation groups with complementary characteristics, allowing us to see higher intra-group homogeneity and higher heterogeneity between groups. The aim of cluster analysis is to sort different objects into groups (Tabachnick 2001). In this case, hierarchical cluster analysis was used where large clusters were composed of smaller clusters (Mefford 1999). Various measures of species richness and diversity were calculated by the mean number of speciesfor all stands. A proximity measure between two plant species was done by measuring their dissimilarity. RelativeEuclidean distance was used as dissimilarity index. Ward’s methodwas linkage methods used to measure dissimilarity between plant species.Principal component analysis was also used to identify plant groups. By adding stand cover types, environmental variables, and soil nutrient variables, we were able to identify plant associations depending on site characteristics.

© Caroline Laberge Pelletier, 2007