Welcome to Online Research .....

Most important, effective research paper collection in the world

Effects of Armillaria root disease on the growth of Picea mariana trees in the boreal plains of central Canada


A.R. WestwoodaCorresponding Author Contact InformationE-mail The Corresponding Author, F. Conciatoria, J.C. Tardifa, K. Knowlesb1
aCentre for Forest Interdisciplinary Research (C-FIR) and Department of Biology, University of Winnipeg, 515 Portage Avenue, Winnipeg, Manitoba, Canada R2B 2E9
bManitoba Conservation, Forestry Branch, 200 Saulteaux Crescent, Winnipeg, Manitoba, Canada R3J 3W3
Received 7 September 2011; revised 1 November 2011; Accepted 2 November 2011. Available online 29 November 2011.

 

Abstract

Armillaria spp. are a complex of fungal pathogens affecting populations of trees worldwide, including upland black spruce (Picea mariana (Mill.) B.S.P.). In central Canada, upland black spruce stands are severely infected with Armillaria root disease, which can kill trees across wide areas. In 2007–2008, infected dead and asymptomatic living trees in 12 infection centers were sampled in each of two regions for growth and mortality analyzes. In 2009, a subset of 10 infected dead and 10 asymptomatic living trees from two sites per region were selected for stem analysis. Dendroecological techniques were used to examine mortality patterns and growth changes prior to mortality. The onset of mortality in affected stands occurred quasi-synchronously across the sampling regions, though differences existed among individual sites within each region. Mortality in all black spruce stands occurred at an average of 96–99 years. Testing of incremental growth ratios indicated that infected trees experienced a sustained decline in basal area and volume increment 5–15 years prior to death, as compared to asymptomatic trees. This significant decline in growth was expressed in overall tree productivity. Comparing logistic regression curves of cumulative basal area, height, and volume growth revealed significant differences between asymptomatic and infected trees, indicating that the infected trees grew more quickly at a younger age than asymptomatic trees. It is speculated that their increased vigor and larger root systems may have predisposed these trees for infection as more root area was available for fungal contact. In upland black spruce forests, Armillaria root disease is accelerating forest succession by breaking up the even-aged post-fire cohort and contributing to the presence of dead wood on the forest floor.

Highlights

► Age at mortality and growth loss were quantified in upland black spruce trees infected with Armillaria root disease. ► In all stands, mortality of infected trees occurred at approximately 90 years of age. ► Infected black spruce trees experienced significant basal area and volume growth loss 5–15 years prior to mortality. ► The true date of tree mortality derived at 1.3 m and below was often underestimated due to the presence of missing rings. ► Increased vigor at a younger age may have predisposed trees to infection.
Keywords: Armillaria; Black spruce; Dendroecology; Stem analysis; Mortality; Root disease

 

 

1. Introduction

Predicting and understanding causes of tree mortality is essential to effectively manage forests for timber production as well as ecosystem and wildlife values (Antos et al., 2007). Members of the fungal genus Armillaria spp. occur in boreal, temperate, and tropical forests worldwide, and Armillaria root disease is known to occur on all continents except Antarctica (DeLong, 1995). In North America, the pathogen is considered a significant agent of biological disturbance (Ayres and Lombardero, 2000). Armillaria root disease is present in all forest regions of Canada, affecting over 200 million hectares (Canadian Forest Service, 2010). It is an important contributor to tree mortality in the forests and has resulted in significant economic losses (Bendel and Rigling, 2008). Incidences of Armillaria infection are becoming more common across Canada, and are thought to be favored by forest management activity that creates stumps (Morrison and Mallet, 1996).
In the Canadian Prairie provinces (Alberta, Saskatchewan, and Manitoba) losses in forest resources due to wood decay agents, such as Armillaria, has been estimated at 3% of gross merchantable volume from 1988 to 1992 (Brandt, 1995). The average onset of decay for all three provinces was a stand age of 80 years (Brandt, 1995). British Columbia experienced annual losses in managed forests from Armillaria alone of 2–3 million m3 (Morrison and Mallet, 1996). In Ontario, older black spruce (median age of 70) have exhibited an 8.4% reduction in merchantable volume due to mortality caused by Armillaria (Whitney, 1978).
In Manitoba, Duck Mountain Provincial Forest (DMPF) and Porcupine Provincial Forest (PPF) contain extensive Armillaria root rot infection in upland black spruce (Picea mariana (B.S.P.) Mill.) stands (Knowles, 2004, Knowles, 2007 and [Epp et al., 2009] ). In one DMPF operating area (Clearwater Creek), it was estimated that 2% of potential merchantable black spruce volume has been lost due to tree mortality, and that the area of infection may double over the next 15 years (Knowles, 2007). Another estimate places merchantable volume losses at 14–45% (Pines in Epp et al., 2009). In one PPF operating area (Schade Lake), root disease has caused volume loss estimated at 8% of total potential volume. Fifteen-year projections suggested that the infected area may more than triple in size (Knowles, 2004).
In addition to an increase in mortality associated with Armillaria, studies have suggested a gradual decline of growth in infected trees. Antos et al. (2007) observed that subalpine fir (Abies lasiocarpa Hooker (Nuttall)) infected by Armillaria ostoyae (currently proposed to change to Armillaria solidipes (Burdsall and Volk, 2008)) displayed a decrease in basal area increment (BAI) in the last 40 years prior to death. A gradual decline in accumulation of volume and basal area was also observed. Mallett and Volney (1999) reported for infected lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) losses in annual volume increment (43%), specific volume increment (32%), and height increment (23%). Greater taper also resulted from the infection as Armillaria was observed to have a more pronounced effect on radial growth than height growth (Mallett and Volney, 1999).
One of the main pathways by which Armillaria affects tree growth and survival is by disrupting water conductivity and nutrition. The upward movement of material through the xylem is blocked, resulting in the mortality of bark, cambium, and wood cells, causing growth loss that may end in death (Shigo and Tippett, 1981 and [Cherubini et al., 2002] ). Cambial mortality in a given sector of the trunk is thought to occur following root mortality associated with Armillaria infection (Cherubini et al., 2002). When acting as a primary pathogen, Armillaria kills vigorous trees. In active disease centers, it tends to kill trees between the ages of five and 20 years old in just a few years. It kills 80–100 years old trees in 10–20 years (Mallett, 1992 and [Morrison and Mallet, 1996] ). As an opportunistic secondary pathogen, it often attacks trees that are stressed, weakened, or wounded by adverse environmental conditions, insects, or diseases ( [Piercey-Normore and Bérubé, 2000] and [Cherubini et al., 2002] ). Interactions of Armillaria with other stressors also increases the risk of mortality (Antos et al., 2007).
Given the growing concerns with Armillaria root rot infection in the DMPF and PPF of Manitoba, this study aimed at better documenting the ecological role and impacts of the disease at both the stand and tree levels. The main objectives were to assess (1) the synchronicity of the onset of mortality in black spruce trees across 12 sampling sites distributed in two regions and (2) examine the effects of Armillaria infection on the growth of black spruce trees. This second objective was achieved by tracing the life histories of infected trees and comparing them to that of asymptomatic (no visible evidence of infection) trees. It was hypothesized that mortality would be synchronous within each region (based on similarities between the two areas with regard to climate, soil type, fire history, stand composition, and age), as well as between the two regions. It was also hypothesized that infected trees would show a reduction in basal area, height and volume increment as compared to asymptomatic trees. It was predicted that this decline would result in a decrease in the overall productivity of the infected trees as expressed by cumulative basal area, height, and volume growth curves.

2. Materials and methods

2.1. Study area

The study was conducted in the boreal plains of Manitoba, central Canada (Fig. 1). Two regions located in western Manitoba were sampled; the Duck Mountain Provincial Forest (DMPF, 51°41′ N, 101°01′ W) and Porcupine Provincial Forest (PPF, 52°40′ N, 101°22′ W). The DMPF covers 376,000 ha, enclosing Duck Mountain Provincial Park, which covers 142,200 ha (Epp et al., 2009). Located to the north of the DMPF, the PPF covers 209,000 ha (Government of Manitoba, 2000). Both provincial forests are situated on the Manitoba Escarpment, a geological formation which forms the boundary between the Manitoba Lowlands and the Saskatchewan Plateau. The study area is also part of the boreal plains ecozone, a transition area between northern boreal forest and southern aspen parkland (Smith et al., 1998). Altitude varies widely, with the eastern escarpment of the DMPF rising 300–400 m in elevation. Upland areas consist of mixed-wood deciduous/coniferous forest. Dominant tree species trembling aspen (Populus tremuloides Michx.), white spruce (Picea glauca (Moench) Voss), balsam poplar (Populus balsamifera L.), black spruce, paper birch (Betula papyrifera Marsh.), jack pine (Pinus banksiana Lamb), balsam fir (Abies balsamea (L.) Mill.) and tamarack (Larix laricina (DuRoi) K. Koch) (Tardif, 2004).

Fig. 1.
Location of the study area. Rectangle indicates the general area enclosing both the Duck Mountain Provincial Forest and Porcupine Provincial Forest. Right insets specify sampling sites (transects) in each region. Of the six disease centers in both regions located on the map, L2 and L3 indicate locations where stem analysis was conducted.
The climate of the region is mid-boreal with short, cool summers and cold winters (Girardin and Tardif, 2005). The nearest weather station is located in Swan River (52°03′ N, 101°13′ W, elevation of 347 m), which is situated between the PPF and DMPF. The area features warm summer and cold winters. For the reference period 1971–2000, the average temperature was −18.2 °C in January and the average temperature was 18.1 °C in July. Overall yearly average temperature was 1.6 °C, and average yearly precipitation was 530.3 mm, of which 394.1 mm fell as rain (Environment Canada, 2009).

2.2. Field methods

Initial sampling was conducted in 2007 in upland black spruce stands in the DMPF. Stands were generally dense with evidence of windthrow and downed wood. Balsam fir occurred in the understory, as well as some scattered paper birch, jack pine, and beaked hazel (Corylus cornuta Marsh). Undergrowth observations consisted of carpet of sphagnum and feather mosses, with common herbaceous species including Rubus idaeus Blanco, Lycopodium spp., Ledum groenlandicum Oeder, Viburnum spp., Cornus canadensis L., Vaccinium spp., and Equisetum spp. Six linear transects (5 × 100m) were established in the Clearwater Creek operating area along known root disease centers (identified by an inventory cruise carried out 5 years prior as well as from aerial reconnaissance flights) (Knowles, 2004 and Knowles, 2007). Along each transect, two increment cores at diameter at breast height (DBH, 1.3 m) were removed from every 10th living black spruce for a total of 144 cores. Two disks were also removed from the stem (one at the base and one at DBH) of each dead tree (standing or fallen) encountered along the transect, for a total of 392 cross sections. Data were recorded concerning the status and condition of the dead trees. All of the sampled dead trees were infected with Armillaria, and exhibited evidence of root and butt rot. In 2008, the same protocol was followed in the PPF and a total of 138 cores and 331 cross sections from black spruce were collected in the Rice Creek operating area
In 2009, two transects in each region were revisited and site characteristics and species present were assessed. Transects 2 and 3 from each region were selected based on accessibility, average age, and absence of evidence of other disturbances (such as extended windthrow). Along each transect, five living trees with no visible symptoms of infection and five infected dead trees were selected for stem analysis (henceforth to be referred to as ‘asymptomatic’ and ‘infected’). Transects were divided into five 20 m segments, with one infected and one asymptomatic tree selected from each zone. Infected trees were chosen from stems that had been previously sampled by Manitoba Conservation in 2007/2008, including only those stems exempt of major rot or missing stem sections. Each infected tree was paired with an asymptomatic control tree sampled within the same zone and with a DBH within 1 cm of that of the infected tree to minimize size variation. Selected asymptomatic trees were at least 10.5 m from the transect (though the majority were over 20 m away). Only living trees showing no visible signs of infection were sampled. This was verified by a visual assessment for the symptoms of Armillaria root disease, completed by examining the tree for crown dieback, chlorosis, cracking and free-flowing sap in the bark near the base, white mycelial fans (under the bark both above, at, and below the root collar), as well as the presence of mycorrhizae in the surrounding soil (Hagle, 2006).
Stem analysis was conducted following Epp and Tardif (2004). Trees were felled and marked lengthwise with paint to indicate an orientation line. Perpendicular marks in a different color were made along the trunk at the base (0 m), 0.5, 1.3 and 2 m, and each subsequent 0.5 m until stem diameter was less than 2 cm. Total tree height was recorded and a cross-section disk removed at each marker. Site, tree number, and disk height were recorded on the upper side of the cross-section, as well as an orientation line corresponding with the upper side of the felled tree.

2.3. Laboratory analysis

In the laboratory, all samples pertaining to the 2007/2008 seasons were prepared and analyzed using standard dendrochronological procedures. Stem disks and cores were dried, sanded and cross-dated. The pointer year method of cross-dating was used (Yamaguchi, 1991) as well as chronologies previously developed for the region (Tardif, 2004 and [Girardin and Tardif, 2005] ). The date of origin and mortality (infected trees only) of all samples (base and DBH) was determined. A total of 126 living asymptomatic trees in the DMPF and 116 in the PPF were dated to determined year of origin, mortality, and longevity. Of the dead infected trees 338 and 140 cross-sections from the DMPF and PPF, respectively were cross-dated to determine year of origin, mortality, and longevity.
All disks collected in 2009 were prepared for stem analysis using the same procedures as those described above. All disks collected from 0 to 3 m were analyzed, as well as disks in 1 m increments above this (e.g. 4, 5, 6 m, etc.). Disks were dated and cross-dated along four radii and both origin and mortality dates were determined at each height. Disks were then scanned and the width of each ring was measured electronically using the program WinDENDRO™ v. 2009a (Régent Instruments Inc., 2009). Measurements were conducted along four radial paths separated by right angles. The radial paths were anchored along the orientation line established at time of tree felling. A number of cross sections could not be measured digitally due to extensive decay or suppression, and were measured on a Velmex measuring stage. All measurements, both digital and manual, were independently validated.

2.4. Statistical analysis

2.4.1. Mortality, origin, and longevity
The origin and mortality date determined from samples taken at DBH were used to analyze tree longevity for each transect. The older of the two cores per tree was used. Longevity was simply determined by subtracting the origin date from the mortality date (the last observed year). Due to non-normality and/or homogeneity of variances issues, the differences in the origin date, mortality date, and longevity of trees was compared between the DMPF and PPF using a Mann–Whitney U-test (n = 6 per region). Within regions, a Kruskal–Wallis test was used to compare the six replicates. If the test revealed a significant difference, Mann–Whitney U-tests were conducted to compare individual transects pairwise in that region. All statistical testing was conducted using SPSS Statistics 17.0 (SPSS Inc., 2008). Origin and mortality dates of infected trees were recorded separately at the base of the tree and DBH for the 2007/2008 samples, and at the height of each disk for the 2009 samples for later analysis.
2.4.2. Stem analysis
Ring-width measurements generated for the 20 asymptomatic and 20 infected trees sampled for stem analysis were converted into annual measurements of basal area, height, and volume. The procedure followed that used in Epp and Tardif (2004). Mean radial growth was calculated using the quadratic mean method. For each of the three growth variables, XLStem™ 1.3a (Régent Instruments Inc., 1999) was used to calculate mean annual increment, with the linear interpolation method being selected for height. This analysis yielded annual increment measures for basal area, height, and volume growth.
To assess if samples from the asymptomatic and infected trees in each region (DMPF and PPF) could be pooled into two groups when comparing stem analysis results, field data on longevity, height, and DBH were compared using Kruskal–Wallis tests. It was determined that the data could be pooled into two groups; asymptomatic trees (N = 20) and infected trees (N = 20) following the absence of any statistical differences (p > 0.05) in these parameters between the regions.
2.4.3. Growth ratio
For each of these variables and for asymptomatic and infected trees, moving growth ratios were calculated starting in the year 1985, which corresponded to the fact that reported mortality can occur 10–20 years after Armillaria infection (Mallett, 1992 and [Morrison and Mallet, 1996] ). The rationale behind comparing growth ratio was that in absence of infection, both groups of trees should have experienced similar ratios, suggesting that climate and geometric growth trends may be the main factors influencing growth. Infection by Armillaria would be expected to cause the growth ratio of the two groups to diverge. The moving growth ratio for each year for each tree was calculated using the formula:View the MathML source
Where MRj is the moving ratio for year j (e.g. 1985), gf-j is the average incremental growth from year f to year j (1981–1985), and ga–e is the average incremental growth from year a to e (1976–1980). A moving ratio value was calculated for each tree, for every year from 1985 to the end of the tree’s life. The moving ratios for the 20 infected trees and the 20 asymptomatic trees were then tested for each year for significant differences using a Mann–Whitney U-test (α = 0.05). This was repeated for each year until 2003, when the number of observations became too low to conduct the test. The same procedure was used for basal area, height, and volume ratios.
To determine if the overall cumulative growth (the summation of annual increments over time) for basal area, height, and volume differed among dead and living trees, a 3-parameter logistic regression model was used:View the MathML source
where y is the value of the growth variable, x is the year of increment, x0 is the inflection point of the curve (indicating the year at which 50% of growth is reached), a is the asymptotic level of growth, and b is the growth constant (slope). Regression curves for basal area, height, and volume were calculated by SigmaPlot 9.01 (Systat Software Inc., 2004), using the average cumulative growth for each year. A model was calculated for the living trees (N = 20) as well as the dead trees (N = 20). Following this, 3-parameter sigmoidal regression models were generated for basal area, height, and volume, for each individual tree. From these individual models (a total of 60), the coefficients a, b, and x0 for the living trees were tested against the coefficients of the dead trees with Mann–Whitney U-tests (α = 0.05) for statistical differences.

3. Results

3.1. Origin, mortality, and longevity of black spruce trees

At the regional level, origin dates (as measured at DBH) were not significantly different between the DMPF and PPF (Mann–Whitney U-test: p = 0.818, n = 12) the average origin date of the stand in each region being the year 1902 (Table 1). Within the DMPF, origin dates also did not significantly differ among the 6 transects (Kruskal–Wallis test: p = 0.066, n = 6). Significant differences were however observed among the six transects in the PPF (Kruskal–Wallis test: p < 0.001, n = 6). For example, trees in transect 3 had an average origin data of 1906, which was significantly later than other transects (Table 1). Trees in transect 5 had an earlier mean origin date (1896), which was essentially due to the presence of remnant trees with origin dates prior to the 1890s (not presented).
Table 1. Summary statistics of dead tree data collected from diameter at breast height in 2007–2008. Different letters (a and b) indicates a significance difference between the group of that letter and the group of another letter, based on a systematic pairwise series of Mann–Whitney U-tests (α < 0.05). Pairwise comparisons were completed among transects 1–6 within each region. Only cross-sections collected from dead black spruce that included the pith date were used for the analyzes.
Transect
Origin date
Mortality date
Longevity
NMeanStd devMeanStd devMeanStd dev
DuckMountain
Transect 1301902a22000a1100a5
Transect 2161903a22001a298a9
Transect 3451901a<12000a199a6
Transect 4331902a11999a196a8
Transect 5291903a12000a197a7
Transect 6161901a<12004b1104b4
All Transects61902120012997
PorcupineMountain
Transect 1201903b11998a295ab8
Transect 2291901a<11997a195ab8
Transect 3331906c<11997a191a8
Transect 4201901a12003b1102c6
Transect 5191896ab31998a2101bc15
Transect 6191902ab11999a296b8
All Transects619023199929610
While all stands selected (except one) originated from the same time period, mortality dates were not synchronous between the DMPF and PPF (Mann–Whitney U-test p = 0.041, n = 12). Mean mortality date in the PPF (1999) occurred an average of 2 years earlier than in the DMPF (2001). A significant difference was observed in mean mortality dates within the DMPF (Kruskal–Wallis test: p = 0.003, n = 6) as trees in transect 6 experienced mortality 3–5 years later than in other transects. Similar results were obtained in the PPF (Kruskal–Wallis test: p = 0.012, n = 6) with trees in transect 4 having died on average 4–6 years later than the other transects (Table 1).
Comparing the two regions for tree longevity (time between tree origin and mortality date), the mean age at which trees died did not statistically differ (Mann–Whitney U-test: p = 0.240, n = 12) with tree death occurring between 96 and 99 years of age (Table 1). Significant differences were however observed among transects within each region (Kruskal–Wallis: p = 0.003, n = 6 for the DMPF, and Kruskal–Wallis: p < 0.000, n = 6 for the PPF). In the DMPF, trees from transect 6 had a statistically significantly greater longevity i.e., 4–7 years greater than in the other transects. In the PPF, variability was observed with longevity of trees ranging from 91 to 102 years (Table 1).

3.2. Mortality date along the stem

The analysis of the date of mortality (last ring produced) recorded for each disk along the stem revealed a lack of consistency in many trees. The date of the last, or most recent, tree ring produced along the entire stem and considered to be the true year of death did not always correspond to the last ring produced on the cross-sections collected at 0, 0.5, and 1.3 m. This indicates that tree rings ceased to be produced in the lower portion of the stem, though the tree remained living and produced tree rings in higher portions of the stem (Fig. 2). On average, mortality dates derived from lower portions in the stem underestimated the true year of death, with an average underestimation of 1.7 years at 0 m, 1.0 year at 0.5 m, and of 0.8 year at 1.3 m.

Fig. 2.
The relationship between year of death determined at a given height as compared to the true year of death (year of the last ring observed over the entire stem) for 0.0 m (A) and 1.3 m (B). Plotted values indicate the observed year of death at that height as compared to true mortality date determined from the dead black spruce trees sampled for stem analysis. The gray line represents linear regression of observed values for illustrative purposes, whereas the black line represents a 1:1 relationship, which would occur if observed year of death was identical to true year of death.

3.3. Growth ratios between infected and asymptomatic trees

To quantify a potential growth loss in trees after infection by Armillaria and prior to mortality, growth ratios were calculated for each asymptomatic and infected tree using annual increment in basal area, height and volume. Prior to any detrimental impact of Armillaria infection on tree growth, little differences in growth ratios were observed between the two groups of trees (Fig. 3). However, prior to mortality, infected trees recorded significant and systematic decline in basal area increment compared to asymptomatic trees (Fig. 3a). This decline in basal area increment ratio was observed from 1997 onward indicating greater basal area decrease in infected trees when comparing the ratio of the basal area growth for the period 1993–1997 to that of 1989–1992. On average, the basal area increment growth ratio of the infected trees was on average 19.0% less than that of the asymptomatic trees from 1997 to 2003. In contrast to basal area, no significant differences were observed in the height increment ratios between asymptomatic and infected trees (Fig. 3b). Similar to basal area, a significant and systematic decline in volume increment growth ratios of infected trees compared to asymptomatic trees was observed from 1997 onward (Fig. 3c). The volume increment growth ratio of infected trees was on average 23.1% less than that of the asymptomatic trees from 1997 to 2003. Volume ratios were also significantly less in infected trees in years 1988 and 1989 (Fig. 3c).

Fig. 3.
Increment growth ratios of 20 infected (black) and 20 asymptomatic (gray) trees over time where (A) basal area (B) height and (C) volume. Each year represents the last year of the ten year testing period. Numbers represent Mann–Whiney U-test P values, where the bold values and symbol indicates a significant difference (α < 0.05). Testing ended after 2003 due to low sample size. Error bars represent one standard error from the mean.

3.4. Cumulative growth between infected and asymptomatic trees

The logistic regressions used to model basal area cumulative increment of both infected and asymptomatic trees were highly significant (p < 0.001) with respective R-square values of 0.71 and 0.68 (Fig. 4a). Of the three parameters of logistic regression (maximum growth rate, maximum cumulative growth, and time at which maximum growth rate was reached), growth rate and maximum cumulative growth did not differ significantly (p > 0.05) between asymptomatic and infected trees (Table 2). However the time at which maximum growth rate was reached was significantly different, and occurred 10–18 years earlier in infected trees than in asymptomatic trees (Table 2). Regression of cumulative height increment yielded a model that explained most of the variation in both asymptomatic (R2 = 0.92, p < 0.001) and infected (R2 = 0.94, p < 0.001) trees (Fig. 4b) indicating that height growth was much less variable among trees than basal area or volume growth (Fig. 4a). All three height growth parameters differed significantly between asymptomatic and infected trees (Table 2). The maximum cumulative height of asymptomatic trees was 1.16 m higher than infected trees, and the maximum growth rate of asymptomatic trees was higher than that of infected trees. The infected trees however reached their maximum growth rate 8 years earlier than the asymptomatic trees. Regressions of cumulative volume increment (Fig. 4c) were also highly significant for both asymptomatic (R2 = 0.66, p < 0.001) and infected (R2 = 0.66, p < 0.001) trees. The maximum cumulative volume did not differ between the two groups. Maximum growth rate did significantly differ, with asymptomatic trees having a higher growth rate than the infected trees. The time at which maximum growth was reached was significantly different between the two groups, as infected trees reached half of their maximum volume 11.6 years earlier than the asymptomatic trees (Table 2).

Fig. 4.
Three-parameter logistic regression fitted to the cumulative growth of all 20 infected (black) and 20 asymptomatic (gray) trees in the Duck Mountain Provincial Forest and Porcupine Provincial Forest, where: (A) basal area (B) height and (C) volume. Error bars represent standard error of the mean.
Table 2. Mean and standard deviation of three sets of logistic growth parameters derived from each individual models fitted to the 20 asymptomatic and 20 infected trees. Predicted growth parameters of the logistic curve are a (maximum cumulative growth), b (maximum growth rate), and x0 (time at which maximum growth was reached). The symbol and bold font indicates a significant result according to Mann–Whitney U-tests (p < 0.05).


Asymptomatic trees
Infected trees


NminmaxMeanStandard deviationNminmaxMeanStandard deviationP
BA (cm2)a2070.64376.94177.5876.182071.32328.25180.2671.810.871

b2012.0321.7616.932.722011.0127.2516.033.760.204

X02047.1879.6564.208.312019.5766.9654.0210.460.002
Height (m)a2010.8720.8416.202.172011.5518.5615.041.760.033

b2014.1622.0618.562.142011.0926.4416.483.430.009

X02040.0559.4349.415.572028.5348.7141.405.700.000
Volume (dm3)a2044.25467.76179.53101.252057.62333.01171.3079.620.957

b2012.8031.1918.234.10209.9722.1215.362.890.015

X02058.26107.8680.7112.472052.8282.4069.148.740.002

4. Discussion

4.1. Black spruce mortality

Mean mortality date between the DMPF and PPF differed by 3 years; black spruce having an average longevity between 96 and 99 years. Given that Armillaria usually kills mature trees in 10–20 years (Mallett, 1992 and [Morrison and Mallet, 1996] ), it is likely that that infection in both regions may have occurred at roughly 80 years of age. Given the potential error in accurately determining the origin date of black spruce trees associated with the difficulty of sampling between the root collar and height of the first year seedling (DesRochers and Gagnon, 1997), as well as the error resulting from determining mortality date using samples from DBH, we assume that the observed differences in longevity among black spruce trees between regions were not biologically meaningful.
In both regions, all 12 upland black spruce stands studied were even-aged, with the exception of one stand in the PPF having remnant trees which likely survived previous fires. All stands presumably originated from stand- replacing fires corresponding to a decade of severe drought and fire (1885–1894) in the DMPF (Tardif, 2004). The apparent synchronicity of the onset of infection/mortality in black spruce trees across the 12 stands from both regions confirmed our first hypothesis. Armillaria is considered a ‘disease of the site’, as established mycelia are essentially permanent (Hagle, 2006). It naturally exists in the soil, but usually exists in a host-disease equilibrium whereby it is rarely a primary mortality agent, but decomposes dead woody tissues in the soil. When the equilibrium is compromised, active disease centers form, and Armillaria becomes a primary pathogen directly responsible for readily killing trees (Morrison and Mallet, 1996). In older stands, Armillaria operates as an opportunistic secondary pathogen that tends to attack stressed or weakened trees ( [Piercey-Normore and Bérubé, 2000] and [Cherubini et al., 2002] ). It is still unknown if susceptibility to infection resulted from a disturbing event that affected both the DMPF and PPF, or if it was simply associated with trees reaching a certain size or age.

4.2. Black spruce growth

Our results indicated that trees infected by Armillaria registered a continuous and greater decline in basal area and volume growth ratios prior to mortality compared to asymptomatic trees. In contrast, height increment ratios in infected black spruce trees prior to mortality were similar to those of asymptomatic trees thus contradicting our hypothesis. Mallett and Volney (1999), when comparing living infected and living uninfected lodgepole pine, also observed that growth reductions in infected trees were more heavily expressed in basal area increment than height through the use of a growth loss index. Our analyzes showed that growth decline in individual black spruce trees started about 5–15 years prior to mortality; a period shorter than reported in other studies. Antos et al. (2007) noted that subalpine fir (Abies lasiocarpa (Hooker) Nuttall) trees infected with Armillaria presented a gradual and long lasting basal area increment decline (∼40 years) prior to tree mortality compared to living control trees. Bloomberg and Morrison (1989), using a measure of relative tree growth rate over short periods, noted that losses in volume growth were substantial and cumulative over a prolonged period. Cherubini et al. (2002) observed that mountain pine (Pinus mugo Terra) trees infected by Armillaria showed declining radial growth with several decades of suppression but the authors were not able to determine if this effect was caused by the pathogen alone, competition from neighboring trees, or a combination of the two. Cruickshank et al. (2011) noted that Armillaria root disease reduced volume yield of Douglas fir soon after infection, accumulating for up to 40 years with no signs of recovery.
In upland black spruce trees, mortality was found to quickly follow the decline in growth increment ratios by approximately a decade. Although infected trees experienced stress associated with the disease, mortality may have ultimately not been directly caused by the pathogen. One likely cause of mortality is windthrow which entails stem snapping or uprooting from the force of wind. Black spruce trees weakened by Armillaria may thus be more susceptible than healthy ones to windthrow. Soil conditions are not thought to impact the incidence of uprooting in black spruce trees (Elie and Ruel, 2005) but increased risk of windthrow has been associated with stem and root rot (Bergeron et al., 2009). Once Armillaria-infected trees begin to die, gap creation occurs, potentially leading to additional mortality due to windthrow. Death by blow-down of adjacent trees is also known to predispose black spruce trees to windthrow (Taylor and MacLean, 2007).
The tree growth/Armillaria association was also reflected in infected trees reaching a lower achievable maximum height and volume compared to healthy trees. The infected black spruce trees in this study were shown to have been initially more vigorous than non-infected trees as they reached half of their maximum height and volume growth at a significantly earlier time compared to living trees. Bloomberg and Morrison (1989) speculated that vigorous trees were likely to become infected sooner and more extensively due to their larger root systems. Cruickshank (2000) also observed that the incidence of infection increased with tree size. Garbelotto et al. (1997) suggested that factors such as the frequency of root contacts/grafts, root architecture, and root size may affect the efficiency of fungal spread in a similar fungus (Heterobasidion annosum). Since infection is related to the spatial distribution of inoculum (Rosso, 1994), trees with larger root systems presumably have more surface area exposed, and thus a higher chance of encountering fungal rhizomorphs or grafting roots with infected trees. It is thus likely that the initial vigor in infected trees may have predisposed them for infection and eventual mortality associated with Armillaria.
While acknowledging that the decline in growth in black spruce was accentuated by the fungus, it is still uncertain if the attack by the fungus was at the origin of the decline or if other external factors leading to a decline in growth may have been at the origin of the attack. It is possible that external stressors such as frost damages may have subsequently compromised tree vigor and defense mechanisms leading to infection (Cherubini et al., 2002). In our sampled trees, many frost rings and white earlywood rings (tree rings characterized earlywood tracheids having thinner secondary walls than in control tree rings) were observed in both infected and asymptomatic black spruce trees (Tardif et al., 2011). These anomalies would need to be quantified between both groups, sites, and regions to better assess if they could be associated with tree vigor being compromised. It is still unknown if these tree-ring anomalies are encountered in the roots and if so, if they could have had a role in allowing mycelium to proceed along and within infected roots to the root collar resulting in stem girdling and subsequent tree mortality. All these points merit further investigation.

4.3. Locally absent rings in infected black spruce

Stem analysis was crucial in documenting that growth decline in infected black spruce trees was often accompanied by the interruption of radial growth in the lower stem. In black spruce trees, Armillaria was observed to cause a reduction in radial growth in the area immediately above the infected root extending up to 3 m above the soil line (Cruickshank, 2002). Discontinuous or locally absent rings are a result of locally inactive cambia and are common in trees that are dying due to minimally available resources, reduced ability of the crown to assimilate nutrients, or local cambial death (Schweingruber, 2007). It is likely that in infected black spruce trees the absence of ring formation at lower stem was associated with local cambial death caused by stem invasion by fungal hyphae (Shigo and Tippett, 1981). Armillariamellea (Vahl: Fr) was reported to surround the stem with hyphal strands that advance at different speeds (Schweingruber, 2007). In the upper stem, tree growth may have been maintained through the use of older tracheid conduits that are still functional (Cherubini et al., 2002).
Not recognizing the potential presence of locally absent rings in the lower stem (DBH and below) of infected trees makes it difficult to precisely estimate true mortality dates. In most cases, this would lead to an underestimation of the tree age at time of mortality. The non-recognition of missing or locally absent rings would also affect the determination of tree longevity and length of growth decline period, thus influencing all productivity assessments. If the presence of locally absent rings in the lower stem is valid for other diseases and tree species, sampling only at or below DBH may also have undetermined implications when modeling tree mortality.

4.4. Black spruce stand dynamics and management

In boreal forests, upland black spruce stands usually establish as even-aged cohorts following stand-replacing fires (Tardif, 2004). In the absence of recurring fires, gap dynamics associated with pathogens, windthrow and insect outbreaks may become the prevalent disturbance in mature stands (Aakala et al., 2007). In the DMPF, a large proportion of the black spruce stands originated from large fires in the 1880s ( [Gill, 1930] and Tardif, 2004) and these mature forests are currently in transition (Epp et al., 2009). In absence of fire and with time, stand breakup occurs as the post-fire even-aged black spruce cohort experiences mortality and senescence. This process is accompanied by increase complexity in the stand structure and may be accompanied by species composition change (Aakala et al., 2007). This process eventually led to an uneven-age structure with more open canopy (St-Denis et al., 2010). In black spruce stands, break-up of the canopy has been described to occur in a clustered arrangement (Lussier et al., 2002) which has often been attributed to the contagious properties of biotic mortality agents (Aakala et al., 2007). Armillaria has been previously associated with clustered mortality of mountain pine (Dobbertin et al., 2001). The creation of active Armillaria disease centers in the DMPF and PPF is accelerating this process by promoting canopy gaps (Epp et al., 2009).
In this study, the upland black spruce stands are in an area of active forest management and mortality due to Armillaria rot disease has negative economic implications (Knowles, 2007) by its impact, amongst others, on determination of the annual allowable cut (AAC) given the high proportion of stands about 130 years of age. At the moment, an overestimation in the amount of available wood may be expected assuming Armillaria-related mortality is not factored into the AAC modeling process. In Czechoslovakia, Hrib et al. (1983) recommended that harvesting of Armillaria-infected norway spruce (Picea abies (L.) Karst) occur at 70–80 years of age, before significant growth decline and mortality. In our study area, the average commercial rotation age of upland black spruce stands is 75 years (Epp et al., 2009) which seems suitable to avoid growth decline, rot and tree mortality. However, harvesting often occurs well after rotation age and wood supply models currently consider that these forests will be available for future use.
Managing upland black spruce stands in a context of forest fire suppression, changing climate, and aging forests poses a number of challenges. Harvesting may not be the most pragmatic option as Armillaria is considered a ‘disease of the site’ which spreads in expansive mycelial clones which range in size from 1 to 5 hectares in some areas, and 20–965 hectares in others (Hagle, 2006). Established mycelia are essentially permanent, even if stump and root removal is undertaken ( [DeLong et al., 2005] and [Vasaitis et al., 2008] ). General forest management operations also seem to favor the fungus ( [Morrison and Mallet, 1996] and Cruickshank, 2000) and planted seedlings are highly susceptible to infection (Piercey-Normore and Bérubé, 2000). There is debate as to whether increasing vigor of existing or replanted trees by thinning or other means is an effective strategy to combat the disease (Rosso, 1994, Hagle, 2006 and [Filip et al., 2009] ). In a context where Canadian boreal forestry is moving towards natural-disturbance-based-management, regulating forest age structure by establishing a 75-year rotation age also poses problems associated with the slow disappearance in the landscape of older black spruce stands typical of stages of development defined by Bergeron et al. (2002) as the second and third structural cohorts. Allowing a proportion of infected stands to persist would facilitate the development of subsequent cohorts and structural variability across the landscape (Epp et al., 2009). Lengthening rotation ages (the stand age at which harvesting can commence) would allow for stands to enter second and third structural cohorts, providing for natural variability on the landscape as well as inputs of coarse woody debris that are not possible with traditional silviculture (Epp et al., 2009). Allowing a proportion of Armillaria infected stands to persist would aid in the provision of variable forest habitats that would be useful in managing for biodiversity conservation in the area. For example, ongoing research is suggesting that Armillaria root disease contributes to increased species diversity in infected red pine stands (Ostry and Moore, 2008).

5. Conclusion

Results indicated that the onset of infection of black spruce trees by Armillaria root disease in the two studied regions was fairly synchronous, with mortality observed in trees about 96–99 years of age. Stem analysis revealed that estimating the date of mortality using samples from the base of the tree to 1.3 m DBH yield a slight underestimation, confirming that prior to mortality many Armillaria-infected black spruce trees had ceased to produce rings in the lower portion of the stem. Given that in dendroecology the conventional sampling height along the stem is often between the base and 1.3 m, the necessity of applying a correction factor may be evaluated, as underestimating mortality date may yield inaccurate estimations of growth losses and mortality processes. The absence of the most exterior rings in slowly dying trees may also have, if not addressed, an impact on our ability to model tree susceptibility to mortality when mortality is not a sudden or random process. In black spruce trees, Armillaria caused a systematic reduction in basal area and volume growth starting 5–15 years prior to mortality. This decline also translated into reduced maximum height and volume. Of particular interest was the finding that increased tree vigor early in life may have predisposed trees to infection by Armillaria due to what may be the development of extensive root systems. In a context leading to natural disturbance-based management, further research is needed to assess the ecological role of Armillaria in shaping the structural characteristics of post-fire cohorts, among others, by contributing snags and coarse woody debris and influencing forest succession.

Acknowledgements

We thank Justin Waito, Karine Grotte, Rob Au, Alisha Carlson, and Leanne Dunne for their assistance in field as well as with laboratory work. We are grateful to the anonymous reviewers for their contribution in improving this manuscript. This research was undertaken, in part, thanks to funding from the Canada Research Chair Program. The Natural Sciences and Engineering Research Council of Canada (NSERC) and the University of Winnipeg also supported this research. We are indebted to Manitoba Conservation for their provision of logistical assets, equipment, and information. A. Westwood also benefited for part of this project from an Undergraduate Student Research Award from NSERC.

References

Aakala et al., 2007 T. Aakala, T. Kuuluvainen, L. De Grandpré and S. Gauthier, Trees dying standing in the northeastern boreal old-growth forests of Quebec: spatial patterns, rates, and temporal variation. For. Ecol. Manage.,  37 1 (2007), pp. 50–61.
Antos et al., 2007 J.A. Antos, R. Parish and G.D. Nigh, Growth patterns prior to mortality of mature Abies lasiocarpa in old-growth subalpine forests of southern British Columbia. For. Ecol. Manage.,  255  (2007), pp. 1568–1574.
Ayres and Lombardero, 2000 M.P. Ayres and M.J. Lombardero, Assessing the consequences of climate change for forest herbivores and pathogens. Sci. Total. Environ.,  262  (2000), pp. 263–286.
Bendel and Rigling, 2008 M. Bendel and D. Rigling, Signs and symptoms associated with Heterobasion annosum and Armillaria ostoyae infection in dead and dying mountain pine (Pinus mugo ssp. Uncinata). For. Pathol.,  38  (2008), pp. 61–72.
Bergeron et al., 2002 Y. Bergeron, A. Leduc, B.D. Harvey and S. Gauthier, Natural fire regime: a guide for sustainable management of the Canadian boreal forest. Silva Fennica,  36  (2002), pp. 81–95.
Bergeron et al., 2009 C. Bergeron, J.-C. Ruel, J.-G. Elie and S.J. Mitchell, Root anchorage and stem strength of black spruce (Picea mariana) trees in regular and irregular stands. Forestry,  82 1 (2009), pp. 29–41.
Bloomberg and Morrison, 1989 W.J. Bloomberg and D.J. Morrison, Relationship of growth reduction in douglas-fir to infection by Armillaria root disease in southeastern British Columbia. Phytopathology,  79 4 (1989), pp. 482–487.
Brandt, 1995 Brandt, J.P., 1995. Forest insect- and disease-caused impacts to tember resources of west-central Canada: 1988–1992. Natural Resources Canada, Canadian Forest Service, Northern Forestry Centre. Northwest Region. Information report NOR-X-341. pp. 36..
Burdsall and Volk, 2008 H.H. Burdsall and T.J. Volk, Armillaria solidipes, an older name for the fungus called Armillaria ostoyae. N. Am. Fungi.,  3 7 (2008), pp. 261–267.
Canadian Forest Service, 2010 Canadian Forest Service, 2010. The State of Canada’s Forests, Annual Report 2010. Natural Resources Canada, Canadian Forest Service, Headquarters, Ottawa, p. 45..
Cherubini et al., 2002 P. Cherubini, G. Fontana, D. Rigling, M. Dobbertin, P. Brang and J.L. Innes, Tree-life history prior to death: two fungal root pathogens affect tree-ring growth differently. J. Ecol.,  90  (2002), pp. 839–850.
Cruickshank, 2000 Cruickshank, M., 2000. Volume loss of Douglas-fir infected with Armillaria ostoyae. In: C. Hollstedt, K. Sutherland, and T. Innes (Eds.). Proceedings, From science to management and back: a science forum for southern interior ecosystems of British Columbia. Southern Interior Forest Extension and Research Partnership, Kamloops, BC, pp. 127–129..
Cruickshank, 2002 M. Cruickshank, Accuracy and precision of measuring cross- sectional area in stem disks of Douglas-fir infected by Armillaria root disease. Can. J. For. Res.,  32  (2002), pp. 1542–1547.
Cruickshank et al., 2011 M.G. Cruickshank, D.J. Morrison and A. Lalumière, Site, plot, and individual tree yield reduction of interior Douglas-fir associated with non-lethal infection by Armillaria root disease in southern British Columbia. For. Ecol. Manage.,  261  (2011), pp. 297–307.
DeLong, 1995 DeLong, D.L., 1995. Geographic distribution of Armillaria species. Nelson Forest Region, Research Summary 015. Ministry of Mines, Forests, and Lands, Government of British Columbia. Available at <http://www.for.gov.bc.ca/hfd/pubs/RSI/FSP/Nelson/RSN015/RSN015.htm>..
DeLong et al., 2005 D.L. DeLong, S.W. Simard, P.G. Comeau, P.R. Dykstra and S.J. Mitchell, Survival and growth response of seedlings in root disease infected partial cuts in the Interior Cedar Hemlock zone of southeastern British Columbia. For. Ecol. Manage.,  206  (2005), pp. 365–379.
DesRochers and Gagnon, 1997 A. DesRochers and R. Gagnon, Is ring count at ground level a good estimation of black spruce age?. Can. J. For. Res.,  27  (1997), pp. 1263–1267.
Dobbertin et al., 2001 M. Dobbertin, A. Baltensweiler and D. Rigling, Tree mortality in an unmanaged mountain pine (Pinus mugo var. Uncinata) stand in the Swiss National Park impacted by root rot fungi. For. Ecol. Manage.,  145  (2001), pp. 79–89.
Elie and Ruel, 2005 J.-G. Elie and J.-C. Ruel, Windthrow hazard modeling in boreal forests of black spruce and jack pine. Can. J. For. Res.,  35  (2005), pp. 2655–2663.
Environment Canada, 2009 Environment Canada, 2009. Canadian climate normals 1971-2000: Swan River, Manitoba. Canada’s National Climate Archive. Environment Canada, Government of Canada. Accessed 10 February 2010. Available at: <http://climate.weatheroffice.gc.ca/climate_normals/results_e.html?Province=ALL&StationName=swan%20river&SearchType=BeginsWith&LocateBy=Province&Proximity=25&ProximityFrom=City&StationNumber=&IDType=MSC&CityName=&ParkName=&LatitudeDegrees=&LatitudeMinutes=&LongitudeDegrees=&LongitudeMinutes=&NormalsClass=A&SelNormals=&StnId=3849&>..
Epp and Tardif, 2004 B. Epp and J.C. Tardif, Effects of Lodgepole pine dwarf mistletoe, Archeuthobium americanum, on jack pine, Pinus banksiana, growth in Manitoba. Can. Field-Nat.,  118 4 (2004), pp. 595–601.
Epp et al., 2009 B. Epp, J.C. Tardif, N.C. Kenkel and L. De Grandpré, Forest dynamics of the duck mountain provincial forest, Manitoba, and the implications for forest management, S. Gauthier, M.-A. Vaillancourt, Alain. Leduc, L. De Grandpré, D. Kneeshaw, H. Morin, P. Drapeau, Y. Bergeron, Editors , Ecosystem Management in the Boreal Forest, Presses de l’Université du Québec, Quebec City (2009).
Filip et al., 2009 G.M. Filip, S.A. Fitzgerald, K.L. Chadwick and T.A. Max, Thinning ponderosa pine affected by Armillaria root disease: 40 years of growth and mortality on an infected site in Central Oregan. West J. Appl. For.,  24 2 (2009), pp. 88–94.
Garbelotto et al., 1997 M. Garbelotto, G. Slaughter, T. Popenuck, F.W. Cobb and T.D. Bruns, Secondary spread of Heterobasidion annosum in white fir root-disease centers. Can. J. For. Res.,  27 5 (1997), pp. 766–773.
Gill, 1930 C.B. Gill, Cyclic forest phenomena. Forest. Chron.,  6  (1930), pp. 42–56.
Girardin and Tardif, 2005 M.P. Girardin and J. Tardif, Sensitivity of tree growth to the atmospheric vertical profile in the Boreal Plains of Manitoba, Canada. Can. J. For. Res.,  35  (2005), pp. 48–64.
Government of Manitoba,, 2000 Government of Manitoba, 2000. Designation of Provincial Forests Regulation, amendment. The Forest Act (C.C.S.M. c. F150). The Queen’s Printer, Province of Manitoba..
Hagle, 2006 Hagle, S.K., 2006. Armillaria root disease: ecology and management, United States Department of Agriculture, Information Report 11.1, Washington, DC..
Hrib et al., 1983 J. Hrib, J. Kyncl and A. Cerny, A tree-ring study of Norway spruce infected with the wood-destroying fungus Armillaria mellea. Eur. J. For. Pathol.,  13  (1983), pp. 160–165.
Knowles, 2004 Knowles, K., 2004. Schade Lake root disease survey. Manitoba Conservation, Forestry Branch, internal report. Winnipeg, Manitoba, p. 12..
Knowles, 2007 Knowles, K., 2007. Clearwater Creek root disease survey, 2005-06. Manitoba Conservation, Forestry Branch, Winnipeg, internal report. Winnipeg, Manitoba, p. 12..
Lussier et al., 2002 J.-M. Lussier, H. Morin and R. Gagnon, Mortality in black spruce stands of fire or clear-cut origin. Can. J. For. Res.,  32  (2002), pp. 539–547.
Mallett, 1992 Mallett, K.I., 1992. Armillaria ostayae root rot in the Canadian prairie provinces. Natural Resources Canada, Canadian Forest Service, Northern Forest Centre. Edmonton, Alberta, Report NOR-X-329, p. 22..
Mallett and Volney, 1999 K.I. Mallett and W.J.A. Volney, The effect of Armillaria root disease on lodgepole pine tree growth. Can. J. For. Res.,  29  (1999), pp. 252–259.
Morrison and Mallet, 1996 D.J. Morrison and K.I. Mallet, Silvicultural management of Armillaria root disease in western Canadian forests. Can. J. Plant. Pathol.,  18  (1996), pp. 194–199.
Ostry and Moore, 2008 M.E. Ostry and M. Moore, Disease impacts in red pine managed to increase stand complexity. Phytopathology,  98 S118 (2008), p. S118.
Piercey-Normore and Bérubé, 2000 M.D. Piercey-Normore and J.A. Bérubé, Artificial inoculation with Armillaria ostoyae in established conifers stressed by defoliation, planting, and thinning in Newfoundland. Can. J. For. Res.,  30  (2000), pp. 1758–1765. | View Record in Scopus | | Full Text via CrossRef | Cited By in Scopus (3)
Régent Instruments Inc., 1999 Régent Instruments Inc., 1999. XLSTEM™ 1.3a. Québec, Qc..
Régent Instruments Inc., 2009 Régent Instruments Inc., 2009. WinDENDRO v.4 2009b. Québec, Qc..
Rosso, 1994 Rosso, P.H., 1994. Tree vigor and the susceptibility of Douglas-fir to Armillaria root disease. Master’s Thesis. Department of Botany and Plant Pathology, Oregon State University, p. 90..
Schweingruber, 2007 F.H. Schweingruber, Wood Structure and Anatomy, Springer-Verlag, Berlin, Heidelberg (2007), p. 279.
Shigo and Tippett, 1981 Shigo, A.L., Tippett, J.T., 1981. Compartmentalization of decayed wood associated with Armillaria mellea in several tree species. United States Department of Agriculture, Forest Service, Research Paper NE- 48, 1–19..
Smith et al., 1998 Smith, R.E., Veldhuis, H., Mills, G.F., Eilers, R.G., Fraser, W.R., Lelyk, G.W., 1998. Terrestrial ecozones, ecoregions, and ecodistricts of Manitoba: An ecological stratification of Manitoba’s natural landscapes. Report and map at 1:1 500 000 scale. Land Resource Unit, Brandon Research Centre, Research Branch, Agriculture and Agri-Food Canada, Winnipeg, Manitoba. Technical Bulletin 1998-9E. SPSS Inc., 2008. SPSS Release 17.0.0. Chicago, IL..
St-Denis et al., 2010 A. St-Denis, D. Kneeshaw and Y. Bergeron, The role of gaps and tree regeneration in the transition from dense to open black spruce stands. For. Ecol. Manage.,  259  (2010), pp. 469–476. Article | PDF (652 K) | | View Record in Scopus | | Cited By in Scopus (1)
Systat Software Inc., 2004 Systat Software Inc., 2004. Sigmaplot for Windows Version 9.01. San Jose, CA..
Tardif, 2004 Tardif, J., 2004. Fire history in the Duck Mountain Provincial Forest, western Manitoba. Sustainable Forest Management Network, Edmonton, Alberta. Project Reports 2003/2004, p. 30..
Tardif et al., 2011 Tardif, J.C., Waito, J., Conciatori, F. 2011. White earlywood rings and frost rings in black spruce trees from Manitoba, central Canada. Eurodendro 2011 Conference Abstract, September 19-23 2011, Engelberg, Switzerland..
Taylor and Maclean, 2007 S.L. Taylor and D.A. Maclean, Spatiotemporal patterns of mortality in declining balsam fir and spruce stands. For. Ecol. Manage.,  253  (2007), pp. 188–201. Article | PDF (1008 K) | | View Record in Scopus | | Cited By in Scopus (10)
Vasaitis et al., 2008 R. Vasaitis, J. Stenlid, I.M. Thomsen, P. Barklund and A. Dahlberg, Stump removal to control root rot in forest stands, a literature study. Silva Fenn.,  42  (2008), pp. 457–483. | View Record in Scopus | | Cited By in Scopus (9)
Whitney, 1978 Whitney, R.D., 1978. Root rot of spruce and balsam fir in northwestern Ontario. I. Damage and implications for forest management. Canadian Forest Service, Great Lakes Forest Centre. Information Report O-X-284..

1 টি মন্তব্য: