Collection Mémoires et thèses électroniques
AccueilÀ proposNous joindre

5 Results

Table des matières

The study was based on 97 conifer-dominated stands: 53 stands were composed principally of black spruce (SbSb), 17 were composed of black spruce–balsam fir (SbBf), 9 were composed of balsam fir-black spruce (BfSb), and 18 stands were composed of balsam fir (BfBf).Dominant trees had a mean height of 16 m, with a height range from 9 to 27 m.Mean canopy openness was 46% and ranged from 10-90 %. Forest humus depth averaged 22 ± 2 cm, and the mor humus was characterized by a high C:N ratio (mean 57 ± 2) and a mean pH of 2.8. Mean pH of the mineral soil was 4.4.Soils are dominantly humo-ferric podzols (Soil Classification Working Group 1998).Particle size class of the C horizon was dominated by sand (mean of 72% sand). Other environmental and soil variables are summarized in table 1.

Black spruce stands (SbSb) can be characterized by important canopy openness, flat ground, coarse soil texture, low nutrient availability, and a younger establishment age (126 years old for SbSb stand, 148 years old for SbBf, 169 years old for BfSb, and 141 years old forBfBf).Balsam fir stands (BfBf) were denser (closed canopy) and found mainly on slopes. Humus pH was less acidic than in black spruce stands, the humus was also richer in total P, in exchangeable bases, and had a lower C:N ratio than that in black spruce stands (Table 2). In general, results showed that canopy openness differed among stand cover types (SbSb, SbBf, BfSb, and BfBf), decreasing with an increase in presence of balsam fir (p= 0.0011, Fig. 7) and showed that cover types with black spruce dominancy are characterized by lower fertility compared to those dominated by balsam fir (Table 2, and fig. 8.a).

Stand cover types in relation to soil fertility and soil chemistry variables are presented ina principal component analysis (PCA) (Fig.8.a). The first two principal components (PC) explained 47.1 % of the variation. PC 1, which explained 29.6 % of the variation,was mainly linked to humus C:N ratio, mineral soil pH and base saturation on the negative side, and humus total P and K, and pH on the positive side. Therefore, the first PC could be interpreted as a gradient of nutrient availability. The second PC, which explained 17.94 % of the variation, was negatively correlated to sand and positively correlated to silt and clay. Therefore, the second PC could be interpreted as a gradient of soil texture. Mineral soil exchangeable bases (individual bases versus exchange capacity) did not have a strongrepresentation on the principal components. This PCA, that combines correlation circle information on soil richness and stand cover types, showed that SbSb is associated with soil characterized by a low nutrient availability; it was presented on the negative side of the PC1. BfBf is associated with a high nutrient availability; it was presented on the positive side of the PC1 (Fig. 8.a).

Relations amongstand cover types andenvironmental characteristics are presented on figure 8.b. The first two PC axes explained 40.2 % of the variation in the environmental characteristics matrix. The first PC represented 21.2 % of the variation. This axis was positively correlated with canopy openness, mean dominant black spruce age and the cosine of the exposure (West-East). Site slope was negatively correlated with the first PC. The first principal component may be interpreted as a gradient of light available to plants. SbSb was located in the area of the ordination associated with greater light availability, whereas BfBf was located on the opposite part of the ordination. The second PC explained 19 % of the variation. Humus depth and drainage were positively correlated with the second PC. Mineral soil depth and canopy openness were negatively correlated with this axis. Thesecond PC may be interpreted as a gradient ofsoilmoistureand soil depth. BfBf was located in the area of the ordination associated with greater humidity and thicker humus, whereas SbSb was located on the opposite part of the ordination, associated with deeper mineral soil and drier soil (Fig. 8.b).

We can see onfigure 8that SbSb and BfBf are separated in space, suggesting that they have different environmental and soil richness requirements in the study region.SbBf and BfSb are situated between the latter two stand cover types, closer to BfBf. However, the length of their respective lines indicates that they are poorly represented on the principal component.

Note:1. Textural class (see textural classe abacus): percentage of the sites with this soil texture. 2. Mean percent of each particle size in soils across the 97 sites.

Notes:ANOVAs were used to compare means between stand cover types.Means followed by the same letter are not statistically different (p> 0.05), Tukey’s test.Stand cover type codes are as following: SbSb, black spruce;SbBf, black spruce/balsam fir; BfSb, balsam fir/black spruce;BfBf, balsam fir.

Stand cover type codes are as following: SbSb, black spruce; SbBf, black spruce/balsam fir; BfSb, balsam fir/black spruce; BfBf, balsam fir.Different letters indicate significant differences between stand cover types, Tukey’s test (α= 0.05).

Only environmental and soil variables estimated to be significantly correlated (p< 0.05) with one of the individual ericad species and stand cover types are indicated. Environmental variable codes are as follows: slope; drainage; humdepth, humus thickness; ageSb, dominant black spruce mean age; cos_expos, cosines of site exposition; canopen, canopy openness; mindepth, mineral soil depth. Stand cover type codes are as following: St-SbSb, black spruce;St-SbBf, black spruce/balsam fir; St-BfSb, balsam fir/black spruce;St-BfBf, balsam fir. Soil variables codes are as follows: CNhum, humus carbon-nitrogen ratio; silt; clay; exchbase, mineral soil exchangeable bases; Khum, humus total potassium; Phum, humus total phosphorus; pHhum, humus pH; sand; pHmin, mineral soil pH basesat, mineral soil base saturation.

A total of 23 species were observed inthe study plots: 6 non-commercial woody species of which 4 are ericaceous species, 7 herbs, 2 species of the fern family, 5 moss species, and 3 lichens (Table 3).Vacciniumspecies(Vaccinium angustifoliumAit.,V. vitis-idaeaL. andV. myrtilloidesMichx.) were the most frequently found ericaceous species, followed byRhododendron groenlandicum(Oeder) Kron and Judd,andKalmia angustifoliaL (76 %, 69 %, and 52 % respectively;n = 97).Of the three most abundant ericads,Rhododendronhad the highest mean cover (15 %; range between 0-75 %),Vacciniumhad a mean cover of 8 %, (range between 0-41 %), andKalmiahad a mean cover of 6 % (range between 0-48 %).

A high cover ofRhododendron , Kalmia,Vacciniumand lichens (Cladinaspp.) was found in the majority of open black spruce stands (25-40 % stand canopy cover; Table 4). However, these species were also present in denser stands, suggesting a potential problem with ericaceous shrubs after canopy removal during harvest (41-60 % stand canopy cover, and 61-80 % stand canopy cover; Table 4).

Note:Means followed by the same letter are not statistically different (p> 0.05), Tukey’s test.

1. No significative difference was identified du to unbalanced observation’s number for SbSb canopy covers.Kalmiastandard error was 3.09 for SbSb (61-80 %), 4.82 for SbSb (41-60 %), and was 2.37 for SbSb (25-40 %).2. No significative difference was identified du to unbalanced observation’s number for BfBf canopy covers; only one site had presence ofKalmia.

Rhododendronpercent cover was explained by four variables (R2 = 0.51,p< .0001, Table 5).Rhododendronpercent cover was higher in open black spruce stand cover type (SbSb; Figs. 9, and 10), while it was also present, with a lower mean cover, in denser black spruce stand cover types (Table 4);Rhododendronpercent cover increased with greater opening of the canopy (Fig. 11). The abundance ofRhododendronpercent cover was also explained by the presence ofSphagnum, while silt loam soil texture was negatively correlated with cover ofRhododendron.

Note:For each ericad species, explicative variables are enumerated in order of decreasing importance.

All explicative variables are continuous except for stand cover types that are discreet.

Stand cover type codes are as following: SbSb, black spruce; SbBf, black spruce/balsam fir; BfSb, balsam fir/black spruce; BfBf, balsam fir. Different letters indicate significant differences between stand cover types, Tukey’s test(α= 0.05).

Six variables best explained percent cover ofVaccinium(R2 = 0.58,p< .0001, Table 5). The abundace ofVacciniumspecies was explained by the openness of the canopy (Fig. 11) and the presence of a sand texture. On the other hand, when the cover ofEquisetum,Lycopodiumand balsam fir increase, the abundance ofVacciniumdecreases. As observed forRhododendron,Vacciniumwas found mainly in open SbSb stands (Figs. 9, and 10).

Kalmiapercentcover was explained by six variables (R2 = 0.63,p< .0001, Table 5).Kalmiawas found on sites wereRhododendronandCladina stellariswere present.Kalmiacover (12.5 %) was higher in 90 yr-old stands,and decreased with standage.Kalmiawas more commonly found associated with flat topography and coarse soil texture.

Individual ericad species presence in relation to soil fertility and soil chemistry variables is presented in the principal component analysis (PCA) (Fig. 12). The first two principal components (PC) explained 47.1 % of the variation in species presence. PC 1, which explained 29.16 % of the variation, was mainly linked to humus C:N ratio, mineral soil pH and base saturation on the negative side, and humus total P and K, and pH on the positive side. Therefore, the first PC could be interpreted as a gradient of nutrient availability. The second PC, which explained 17.94 % of the variation, was negatively correlated to sand and positively correlated to silt and clay. Therefore, the second PC could be interpreted as a gradient of soil texture. Mineral soil exchangeable bases (individual bases versus exchange capacity) did not have a strongrepresentation on the principal components. This PCA, that combines correlation circle information on soil richness and ericaceous species, showed that ericaceous species are associated with soil characterized by high humus C:N ratio, low humus pH, a coarse soil texture, and low nutrient availability (Fig. 12).Rhododendron, Vaccinium,andKalmiawere situated on the same side as SbSb (Fig. 8.a), showing that these ericads were found in a similar environment.

Only soil variables estimated to be significantly correlated (p < 0.05) with one of the individual ericad species are indicated. Soil variable codes are as following: CNhum, humus carbon-nitrogen ratio; silt; clay; exchbase, mineral soil exchangeable bases; Khum, humus total potassium; pHhum, humus pH; Phum, humus total phosphorus; sand; pHmin, mineral soil pH; basesat, mineral soil base saturation.

Only soil variables estimated to be significantly correlated (p< 0.05) with one of the individual ericad species are indicated. Soil variable codes are as following: CNhum, humus carbon-nitrogen ratio; silt; clay; exchbase, mineral soil exchangeable bases; Khum, humus total potassium; pHhum, humus pH; Phum, humus total phosphorus; sand; pHmin, mineral soil pH; basesat, mineral soil base saturation.

An additional perspective of relations among individual ericad species presence and their soil requirements is presented by a canonical correlation analysis (Fig. 13). This graphic allows one to differentiate individual ericad species in term of soil requirements.Soil variables estimated to besignificantly correlated (p< 0.05) with one of the individual ericad species are indicated. The first axis was positively correlated with silt and clay (fine soil texture) and was negatively correlated with sand (coarse soil texture). The first axis may be interpreted as a gradient of soil texture. The second axis was positively correlated with mineral soil pH, humus total P and K, and was negatively correlated with base saturation, humus pH, and exchangeable bases. The second axis may be interpreted as a gradient of base cation availability.Kalmiawas found on the negative side of both axes.Kalmiacover was positively associated with percent of sand in mineral soil, higher base saturation and relatively higher mineral soil pH. It was negatively associated with finer soil textures (silt and clay), and with greater base cation availability.Vacciniumwas found on the positive side of both axes. It was positively associated with finer soil textures, and with humus total P and K.Rhododendronwas on the positive side of the first axis and on the negative side of the second axis. Its presence was positively associated with the finest soil texture (clay), high exchangeable bases, higher humus pH, and lower mineral soil pH. The humus C:N ratio was poorly represented on the figure, indicating that this variable could not differentiate the three ericad species; all three species are found on sites with high humus C:N rate.

Individual ericad species, and soil variables estimated to be significantly correlated (p < 0.05) with one of the individual ericad species are indicated. Individual ericad species codes are as following: Kalmiaa, Kalmia angustifolia L.; Rhodog, Rhododendron groenlandicum (Oder) Kron and Judd.; Vaccsp , Vaccinium spp. Soil nutrient codes are as following: pHmin, mineral soil pH; Phum, humus total phosphorus; Khum, humus total potassium; silt; clay; pHhum, humus pH; exchbase, exchangeable bases; basesat, base saturation; CNhum, humus C:N ratio; sand.

Individual ericad species, and soil variables estimated to besignificantly correlated (p< 0.05) with one of the individual ericad species are indicated.Individual ericad speciescodes are as following:Kalmiaa,Kalmia angustifoliaL.;Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp , Vacciniumspp.Soil nutrient codes are as following: pHmin, mineral soil pH; Phum, humus total phosphorus; Khum, humus total potassium; silt; clay; pHhum, humus pH; exchbase, exchangeable bases; basesat, base saturation; CNhum, humus C:N ratio; sand.

Relations among environmental characteristics and individual ericad shrub species presence are presented on figure 14. The first two PC axes explained 40.2 % of the variation in the environmental characteristics matrix. The first PC represented 21.2 % of the variation. This axis was positively correlated with canopy openness, mean dominant black spruce age and the cosine of the exposure (West-East). Site slope was negatively correlated with the first PC. The first principal component may be interpreted as a gradient of light available to plants. The ericaceous species were located in the area of the ordination associated with greater light availability. Stands with black spruce presence were also located in the same area, whereas balsam fir stands were located on the opposite part of the ordination (Fig. 8.b). The second PC explained 19 % of the variation. Humus depth and drainage were positively correlated with the second PC. Mineral soil depth and canopy openness were negatively correlated with this axis. Thesecond PC may be interpreted as a gradient ofsoilmoistureand soil depth.

The three individual ericad species are found in a similar environment (Fig.14). Nevertheless, we can see thatRhododendronwas on the upper side of the second PC, whereasVacciniumandKalmiaare on the lower side of the second PC.Rhododendronseems to tolerate more humid soil drainage conditions compared to the other two species.Rhododendronpercent cover was higher on the extremes of the drainage gradient, that is, either dry or humid soil (28 % for class 1, and 28 % for class 6),Vacciniumpercent cover was well distributed among all drainage class, but slightly higher on dry sites (11 % for class 0), andKalmiapercent cover was higher on dry sites (20 % for class 0, and 32 % for 1).VacciniumandKalmiaappear to have higher needs in terms of light availability and mineral soil depth than was observed forRhododendron. Also, the three individual ericad species are negatively correlated with site slope.

Only environmental variables estimated to be significantly correlated (p < 0.05) with one of the individual ericad species are indicated. Environmental variable codes are as following: slope; drainage; humdepth, humus thickness; ageSB, dominant black spruce mean age; cos_expos, cosines of the exposition (west to east); canopen, canopy openness; mindepth, mineral soil depth.

Only environmental variables estimated to be significantly correlated (p< 0.05) with one of the individual ericad species are indicated. Environmental variable codes are as following: slope; drainage; humdepth, humus thickness; ageSB, dominant black spruce mean age; cos_expos, cosines of the exposition (west to east); canopen, canopy openness; mindepth, mineral soil depth.

Relations among environmental characteristics and individual ericad shrub species presence are shown on figure 15. Individual ericad species in this case are associated with specific environmental variables. On the negative side of the first axis,Kalmiawas positively associated with an eastern exposition and negatively associated with poor drainage, steeper slopes and black spruce age. On the positive side of the first axis,Rhododendronwas positively correlated with humus depth, black spruce age, and soil moisture (slow drainage) contrary toKalmia.Vacciniumwas also represented on the positive side of the first axis, but on the negative side of the second axis. This ericad was positively associated with greater mineral soil depth, tolerated steeper slopes and greater soil humidity thanKalmia. The canopy openness was poorly represented on this figure, indicating that this variable could not distinguish individual ericad species. All three species had a high affinity with canopy openness.

Individual ericad species, and environmental variables estimated to be significantly correlated (p < 0.05) with one of the individual ericad species are indicated. Individual ericad species codes are as following: Kalmiaa, Kalmia angustifolia L.; Rhodog, Rhododendron groenlandicum (Oder) Kron and Judd.; Vaccsp , Vaccinium spp. Environmental variable codes are as following: cos_expos, cosines of the exposition (west to east); humdepth, humus thickness; ageSB, dominant black spruce mean age; drainage; slope; mindepth, mineral soil depth; canopen, canopy openness.

Individual ericad species, and environmental variables estimated to besignificantly correlated (p< 0.05) with one of the individual ericad species are indicated.Individual ericad speciescodes are as following:Kalmiaa,Kalmia angustifoliaL.;Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp , Vacciniumspp. Environmental variable codes are as following: cos_expos, cosines of the exposition (west to east); humdepth, humus thickness; ageSB, dominant black spruce mean age; drainage; slope; mindepth, mineral soil depth; canopen, canopy openness.

Individual ericad species distribution in relation to black spuce and balsam fir regeneration and black spruce growth parameters is presented on figure 16. The first two principal components explained 61.9 % of the variation in the ericaceous presence in relation to tree regeneration and growth.The first component represented 39.9 % of the variation. Black spruce merchantable tree mean diameter at breast height (DBH), dominant black spruce mean DBH, balsam fir regeneration, and black spruce mean height were positively correlated with this component. The second component represented 22.1% of the variation and Sb regeneration presence was positively correlated with this component. Individual ericad shrub species (Rhododendron, Vaccinium,andKalmia) were on the opposite side of the tree growth variables, indicating a negative association between ericads and black spruce merchantable tree mean DBH, dominant black spruce mean DBH, and black spruce mean height. Concerning the relation between ericads and black spruce regeneration, the figure 16 showed no significant correlation; there is an angle of almost 90° between their vectors. Information onregeneration and growth measures for all stands are presented in table 2.

Individual ericad speciescodes are as following: Kalmiaa,Kalmia angustifoliaL.; Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp , Vacciniumspp.Growth parametercodes are as follows: seedlingSb, regeneration of black spruce seedlings; seedlingBf, regeneration of balsam fir seedlings; DBHSbmt, black spruce merchantable tree diameter at breast height; DBHSb, black spruce diameter at breast height; heightSb, black spruce height.

Kalmiawas clearly negatively correlated to black spruce merchantable tree mean DBH, dominant black spruce mean DBH, balsam fir and black spruce regeneration, and black spruce mean height.Vacciniumwas positively associated with balsam fir regeneration, black spruce merchantable tree mean DBH, and dominant black spruce mean DBH.Rhododendronwas positively associated with black spruce regeneration and black spruce mean height (Fig. 17).

Individual ericad species, andgrowth variablesestimated to besignificantly correlated (p< 0.05) with one of the individual ericad species are indicated.Individual ericad speciescodes are as following:Kalmiaa,Kalmia angustifoliaL.;Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp , Vacciniumspp. Growth parametercodes are as follows: seedlingBf, regeneration of balsam fir seedling; DBHSb, black spruce diameter at breast height; DBHSbmt, black spruce merchantable tree diameter at breast height; heigthSb, black spruce height; seedlingSb, regeneration of black spruce seedlings.

Percent cover ofRhododendron,Vaccinium, andKalmiadiffered among cover types (p< .0001,p< .0001,p= 0.0003, respectively; Tukey’s test; Table 2). Their covers were consistently higher in SbSb (Total of 47 % cover for the three ericaceous shrubs) and decreased with increasing presence of balsam fir; cover decreased within SbBf (total of 15 % cover for the three ericaceous shrubs), BfSb (total of 1 % cover for the three ericaceous shrubs), and BfBf (total of 4 % cover for the three ericaceous shrubs) (Fig. 9).

The typology formulated by the MRNQ for the study region includes twelve types, according to locality, site physical characteristics (soil texture and drainage), forest type, and vegetation composition.Rhododendroncover tended to be higher in types RE10, RE21, RE11, RE20, and RE22, RE25, and was significantly lower on types RE12, RS22, RS25, RS20, and RS2A respectively (p= 0.0003).Vacciniumcover was highest in RE21, RE10, RE11, and RE20, and significantly lower in RS25, RE25, RS22, and RS2A respectively (p< .0001).Finally,Kalmiamean cover was highest in RE11, RE12, and RE10, and significantly lower in RE21, RE25, RS20, RE22, RS22, RS25, and RS2A respectively (p< .0001) (Table 6).

Note:Ecological 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.
Values with the same letter (comparing percent cover among ecological site types) are not significantlydifferent, Tukey’s test(α = 0.05).

The typology recently formulated for the same study region was composed of six types (MSc thesis, Côté 2006) according to species composition, tree state (health), and stand diameter structure.Rhododendronmean percent cover was higher within Types 5, 2, and 1, and significantly lower in Types 4, and 6.Vacciniumpercent cover was higher in Types 5, 6, 2, and 1, and lower in Type 4.Kalmiacover was higher in Type 5 and significantly lower in Type 4 (Table 7).

Notes: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. Values with the same letter (comparing percent cover among ecological site types) are not significantly different, Tukey’s test(α = 0.05).

Type 1was represented by young even-aged SbSb stands originating from fire or other intense perturbations. Type 2 was even-aged SbSb stands that were older than those in Type 1. Few balsam fir were found in this second group. Type 6 consisted of non-regenerated SbBf stands; regeneration was lacking and was composed mainly of balsam fir. Type 3 was uneven-aged black spruce-balsam fir stands (SbBf) dominated by balsam fir regeneration. Type 4 was uneven-aged balsam fir-black spruce stands (BfSb). This type was the oldest and was positively associated with class B density (canopy closed at 61-80 %). Balsam fir regeneration was predominant (MSc thesis, Côté 2006). Type 5 was represented by an open black spruce (SbSb) stands, closely related to density class D (canopy closed at 25-40 %). Most natural gaps were the consequence of natural perturbations rather that tree mortality (Côté 2006).

The mean percent cover ofRhododendron, Vaccinium,andKalmiawithin stand age classes was evaluated (Table 8). Only the percent cover ofRhododendronvaried with age of black spruce stands, showing higher cover with increasing age of this stand cover type (p=0.0115).

Note:Mean percent cover of individual ericad species were evaluated according to age class within black spruce, black spruce/balsam fir, and balsam fir stand cover types.Values with the same letter are not significantly different(comparing percent cover among age classes), Tukey’s test(α = 0.05). BfSb was not included due to low number of observations.

Fourplant groups were identified (Fig. 18.a). The dotted line represents truncation and allows the identification of four plant associations (Fig. 18.ab). The first group (Ericad and lichen group) was composed of ericaceous species (Kalmia, Rhododendron,andVaccinium) and lichen species (Cladina stellaris, C. mitis,andC. rangiferina). The second group (Equisetum group) was composed ofSphagnumspp., Rubus chamaemorus,Equisetumspp., Lycopodiumspp.,andCarexspp. The third group (Moss group) was composed of an ericad:Gaultheria hispidulaand four bryophytes:Ptilium crista-castrensis , Pleurozium schreber i , Dicranumspp.andPolytrichumspp. The fourth group (Herb group) was composed mainly of boreal herb species:Clintonia borealis , Trientalis borealis,Maianthemum canadense , Linnaea borealis , Coptis groenlandica , Oxalis montana , Cornus canadensisand one bryophyte,Hylocomium splendens(Fig. 18.c).

a) Cluster dendogram presents groups of plant species, and then sub-groups. b) The dotted line presents truncation and allows the identification of four homogeneous groups. c) Plant classification in groups. Plant codes are as follows: Kalmiaa,Kalmia angustifoliaL.; Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp, Vacciniumspp.; Cladiste, Cladina stellaris(Opiz) Brodo;Cladiran,Cladina rangiferina(L.) Nyl.;Cladimit, Cladina mitis(Sandst.) Hustich;Rubus, Rubus chamaemorusL.;Carex, Carexspp.;Sphagn, Sphagnumspp.; Equiset,Equisetumspp.;Lycopo, Lycopodiumspp.;Gaulthe, Gaultheria hispidulaL. Mühl.;Pleuro, Pleurozium schreberi(Brid.) Mitt.;Polytric, Polytrichumspp.;Dicran, Dicranumspp.;Ptilium, Ptilium crista-castrensis(Hedw.) De Not.;Cornus, Cornus canadensisL.;Maianth, Maianthemum canadenseDesf..;Oxalis, Oxalis montanaRaf.;Linnaea, Linnaea borealisL.;Clintoni, Clintonia borealis(Ait.) Raf.;Coptis, Coptis groenlandica(Oeder) Fern.;Triental, Trientalis borealisRaf..;Hyloco, Hylocomium splendens(Hedw.) B.S.G.Each group of plants had an identifying color, which is used for their identification on correlation circles (PCA).

Stand distributions among plant groups are presented on figure 19. The first two principal components of the PCA explained 32 % of the variation in plant species. The first PC represents 19.9 % of the variation and the second PC represents 12.1 % of the variation. On the negative side of the first PC, and on the positive side of the second PC, we found theEricad and lichen group. This groupwas associated with SbSb stands. On the positive side of the first PC, and the positive side of the second PC, we found the Equisetum group, located between SbSb and BfBf stands. On the positive side of the first PC and the negative side of the second PC, we found theHerbgroup, which was associated with BfBf stands. The third group (Mossgroup) was situated on the negative side of the first PC and the negative side of the second PC, more closely associated with SbBf stands.

Plant codes are as following: Kalmiaa,Kalmia angustifoliaL.; Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp, Vacciniumspp.;Rubus, Rubus chamaemorusL.;Gaulthe, Gaultheria hispidulaL. Mühl.;Cornus, Cornus canadensisL.;Clintoni, Clintonia borealis(Ait.) Raf.;Carex, Carexspp.;Triental, Trientalis borealisRaf.;Coptis, Coptis groenlandica(Oeder)Fern.;Linnea, Linnaea borealisL.;Maianth, Maianthemum canadenseDesf.;Oxalis, Oxalis montanaRaf.;Ptilium, Ptilium crista-castrensis(Hedw.) De Not.;Hyloco, Hylocomium splendens(Hedw.) B.S.G.;Pleuro, Pleurozium schreberi(Brid.) Mitt.;Polytric, Polytrichumspp.;Dicran, Dicranumspp.;Sphagn, Sphagnumspp.; Equisetum,Equisetumspp.;Lycopo, Lycopodiumspp.;Cladiste, Cladina stellaris(Opiz) Brodo;Cladiran,Cladina rangiferina(L.) Nyl.;Cladimit, Cladina mitis(Sandst.) Hustich. Stand cover type codes are as following: St-SbSb, black spruce type;St-SbBf, black spruce/balsam fir; St-BfSb, balsam fir/black spruce;St-BfBf, balsam fir. Colored ellipses were made by hand and regroup vegetation species in their plant group according to cluster analysis.

Biplots of plant groups and soil texture permitted to see the plant associations with regards to the soil texture (Fig.20). The Ericad and lichen group was associated with sandy soil, theEquisetumgroup with no particular soil texture (organic layer more important) and theMossgroup with fine texture (silt loam soil). TheHerbgroup was most closely associated with intermediate soil textures (loam, sandy loam).

The ordination of species gives rise to clusters corresponding to habitat characteristics (Fig. 21). On the left side of the axis 1, we found the Ericad and lichen group associated with canopy openness, sandy soil,low-fertility soilsand SbSb stands. On the opposite side of axis 1, still at the top of the second axis, we found the Herb groupassociated with relatively higher humus pH, mineral soil with higher CEC, higher exchangeable bases, older stands, a more humid soil, and BfBf stands. Between those two groups, we found the Equisetum group associated with greater humus depth, mineral soil with relatively higher base saturation, exchangeable acidity, and humus phosphorus, situated between SbSb stands and BfBf stands. On the lower side of axis 2, we found the Mossgroup associated with high humus total K, greater site slope, greater silt and clay in the mineral soil, deeper mineral soils, SbBf stands and BfSb stands.

Plant codes are as following:Kalmiaa,Kalmia angustifoliaL.; Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd.;Vaccsp, Vacciniumspp.;Rubus, Rubus chamaemorusL.;Gaulthe, Gaultheria hispidulaL. Mühl.;Cornus, Cornus canadensisL.;Clintoni, Clintonia borealis(Ait.) Raf.;Carex, Carexspp.;Triental, Trientalis borealisRaf.;Coptis, Coptis groenlandica(Oeder)Fern.;Linnea, Linnaea borealisL.;Maianth, Maianthemum canadenseDesf..;Oxalis, Oxalis montanaRaf.;Ptilium, Ptilium crista-castrensis(Hedw.) De Not.;Hyloco, Hylocomium splendens(Hedw.) B.S.G.;Pleuro, Pleurozium schreberi(Brid.) Mitt.;Polytric, Polytrichumspp.;Dicran, Dicranumspp.;Sphagn, Sphagnumspp.; Equisetum,Equisetumspp.;Lycopo, Lycopodiumspp.;Cladiste, Cladina stellaris(Opiz) Brodo;Cladiran,Cladina rangiferina(L.) Nyl.;Cladimit, Cladina mitis(Sandst.) Hustich.Soil texture codes are follows: Soiltext-S, sand; Soiltext-L, loam; Soiltext-SL, loamy sand; Soiltext-LS, sandy loam; Soiltext-LL, silt loam.Colored ellipses were made by hand and regroup vegetation species in their plant group according to the cluster analysis.

Plants, environment, and soil variables estimated to be significantly correlated (p< 0.05) with one of the individual ericad species are indicated. Plant codes are as following:Kalmiaa,Kalmia angustifoliaL.;Rhodog, Rhododendron groenlandicum(Oder) Kron and Judd. ;Vaccsp , Vacciniumspp.; Rubus, Rubus chamaemorusL.;Gaulthe, Gaultheria hispidulaL. Mühl.;Cornus, Cornus canadensisL.;Clintoni, Clintonia borealis(Ait.) Raf.;Carex, Carexspp.;Triental, Trientalis borealisRaf.;Coptis, Coptis groenlandica(Oeder) Fern.;Linnea, Linnaea borealisL.;Maianth, Maianthemum canadenseDesf..;Oxalis, Oxalis montanaRaf.;Ptilium, Ptilium crista-castrensis(Hedw.) De Not.;Hyloco, Hylocomium splendens(Hedw.) B.S.G.;Pleuro, Pleurozium schreberi(Brid.) Mitt.;Polytric, Polytrichumspp.; Dicran, Dicranumspp.;Sphagn, Sphagnumspp.;Equisetu,Equisetumspp.;Lycopo, Lycopodiumspp.;Cladiste, Cladina stellaris(Opiz) Brodo;Cladiran,Cladina rangiferina(L.) Nyl.;Cladimit, Cladina mitis(Sandst.) Hustich. Environmental variable codes are as follows: drainage; humdepth, humus thickness; mindepth, mineral soil depth; ageSb, black spruce age; canopen, canopy openness; slope. Nutrients codes are as following: sand; Phum, humus total phosphorus; pHhum, humus pH; Khum, humus total potassium; clay; exchbase, exchangeable bases; exchacid, mineral soil H+; CEC, cation exchange capacity; silt; CNhum, humus C:N ratio; CNmin, mineral soil C:N ratio; basesat, base saturation; pHmin, mineral soil pH. Stand cover type codes are as following: SbSb, black spruce; SbBf, black spruce/balsam fir; BfSb, balsam fir/black spruce; BfBf, balsam fir. Colored ellipses were made by hand and regroup vegetation species in their plant group according to the cluster analysis.

© Caroline Laberge Pelletier, 2007