Self-thinning dynamics in a balsam fir (Abies balsamea (L.) Mill.) insect-mediated boreal forest chronosequence

Forest Ecology and Management 241 (2007) 295–309 www.elsevier.com/locate/foreco

Self-thinning dynamics in a balsam fir (Abies balsamea (L.) Mill.) insect-mediated boreal forest chronosequence John W. McCarthy *, Gordon Weetman 1 Forest Sciences Department, University of British Columbia, 2424 Main Mall, Vancouver, BC V6T 1Z4, Canada Received 12 August 2006; received in revised form 27 December 2006; accepted 12 January 2007

Abstract Self-thinning dynamics were examined in a natural, 120-year, insect-mediated balsam fir (Abies balsamea (L.) Mill.) chronosequence in the humid boreal forests of western Newfoundland. The well-developed chronosequence representing the classic stages of stand development provided a unique opportunity to quantify balsam fir self-thinning dynamics and to independently test a first approximation stand density management diagram developed for mixed balsam fir-black spruce (Picea mariana (Mill.) B.S.P.) stands in western Newfoundland. A slope of 1.28 for the self-thinning tree volume–density line (as determined by reduced major axis) was significantly different than the theoretical self-thinning slope of 1.5. This compared very well with other self-thinning studies in Abies stands. Older stands judged as non-self-thinning had a shallower slope approaching unity. The mixed fir-spruce stand density management diagram was found to be a good reflection of stand dynamics and may act as an effective operational tool in tree density management. Balsam fir began to self-thin at 75 percentile heights of 7 m, 60 years of age and stem densities approaching 31,000 stems ha1. Balsam fir ceased self-thinning in 90 year-old stands, with tree heights > 15 m and stem densities < 3000 stems ha. The relative accumulation of live and dead basal area across the chronosequence was best explained by the relative contribution of insect herbivory, self-thinning and density-independent mortality to tree death. The relative proportion of dead basal area declined steadily from a high of 80% in the youngest stands to a low of 20% in the 60-year-old stands, increasing to levels not exceeding 30% in the oldest stands. Stand break-up associated with the onset of density-independent tree mortality occurred in stands approaching 90 years of age. # 2007 Elsevier B.V. All rights reserved. Keywords: Newfoundland; Stand development; Reduced major axis; Spruce budworm; Hemlock looper

1. Introduction The boreal forests of Canada are generally of fire origin, but in the moist eastern boreal forests dominated by late succession balsam fir (Abies balsamea (L.) Mill.), forest fires are less frequent (Johnson, 1995; Bergeron et al., 2001). This is particularly so in the ‘‘wet boreal’’ forests of western Newfoundland where fire is rare or non-existent (Meades and Moores, 1994; Thompson et al., 2003). In these generally humid forests of eastern Canada, insect herbivory, particularly by the spruce budworm (Choristoneura fumiferana (Clem.)), hemlock looper (Lambdina fiscellaria fiscellaria Guen.) and balsam fir sawfly (Neodiprion abietis Harris), is the dominant * Corresponding author. Present address: P.O. Box 1238, Guelph, Ont. N1H 6N6, Canada. Tel.: +1 519 824 1250; fax: +1 519 767 0994. E-mail addresses: [email protected] (J.W. McCarthy), [email protected] (G. Weetman). 1 Tel.: +1 604 822 2504; fax: +1 604 822 9102. 0378-1127/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2007.01.001

disturbance factor (Holling, 1992; Engelmark, 1999). Three major spruce budworm outbreaks starting in 1910, 1940 and 1970 affected about 10, 25 and 57 million hectares of forest, respectively (Blais, 1983; Morin and Laprise, 1990). The hemlock looper has caused extensive stand mortality in Newfoundland (Hudak et al., 1996) and to a lesser extent in New Brunswick (MacLean and Ebert, 1999). More recently, historically unprecedented outbreaks of balsam fir sawfly have occurred in western Newfoundland (Moreau, 2006). The humid boreal forest landscape is dominated by stands originating from insect defoliation. The new stands are the result of the vigorous release of a well-established balsam fir seedling bank (Greene et al., 1999). This link between insectmediated stand collapse and vigorous stand re-initiation is so tightly coupled that the balsam fir-spruce budworm ecosystem is best described as a disturbance-based, self-regulating, selfperpetuating system (Baskerville, 1975; MacLean, 1984). It has been extremely difficult to examine the stand dynamics of insect origin stands in a chronosequence due to

296

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

the long history of pulpwood harvesting in balsam fir stands of eastern Canada. Harvesting has been widespread, often damaging residual seedling banks and creating an unnatural and highly variable pattern of second-growth stand development. An area was noted in western Newfoundland where an insect-mediated balsam fir chronosequence could be examined in a landscape that had never been logged. The Little Grand Lake Provisional Ecological Reserve is composed of a finescale mosaic of stands of diverse age, structure and developmental sequence (McCarthy, 2004). This landscape heterogeneity is controlled for the most part by recurrent standreplacing outbreaks of the hemlock looper and spruce budworm (Otvos and Moody, 1978; Otvos et al., 1979). Periodic outbreaks that kill mature canopies and replace them with vigorous young stands have created a natural chronosequence of relatively homogeneous even-aged stands. The landscapelevel mosaic patch structure has provided a unique opportunity to examine the relationship between insect herbivory and stand structure. In most other forest areas of Newfoundland, and indeed throughout much of eastern Canada, such an opportunity does not exist because of the confounding effect of extensive forest harvesting. Balsam fir, given its ability to form persistent, welldeveloped seedling banks under mature canopies, responds vigorously to release after insect-mediated canopy disturbance and undergoes intense self-thinning during stand development. Density-dependent competition during self-thinning is generally considered to follow the 3/2 power law (Yoda et al., 1963). Self-thinning in Abies stands has been traditionally examined in Abies wave forests (Oshima et al., 1958; Kuroiwa, 1959; Tadaki et al., 1977; Mohler et al., 1978; Kohyama and Fujita, 1981; Sprugel and Bormann, 1981; Sprugel, 1984; Moloney, 1986; Kohyama et al., 1990; Robertson, 1993) or in stands that provided a range in tree densities needed for the construction of stand density management diagrams (SDMDs) (Sturtevant et al., 1998; Wilson et al., 1999; Be´gin et al., 2001; Solomon and Zhang, 2002). This study examines self-thinning dynamics within a natural, insect-mediated 120-year chronosequence of balsam fir-dominated stands. The well-established chronosequence provides a unique opportunity to examine three main objectives: (1) the quantification of balsam fir self-thinning dynamics in an insect-mediated natural chronosequence of stands, (2) a comparison of self-thinning dynamics with published accounts of self-thinning in other Abies stands and (3) independent testing of a first approximation SDMD developed for mixed balsam fir-black spruce (Picea mariana (Mill.) B.S.P.) stands of western Newfoundland (Sturtevant et al., 1998). Using established self-thinning theory, it is possible to determine tree heights at which self-thinning begins, the beginning of stand break-up and consequent ‘‘gap dynamics’’ in the absence of insect herbivory, the accumulation of dead wood with stand development and the changing characteristics of stand and stock table development in these natural disturbance origin stands.

2. Materials and methods 2.1. Study area The 106 km2 study area (350–500 m above sea level) is located within the Little Grand Lake Provisional Ecological Reserve in western Newfoundland (488380 N, 578490 W). The study area is considered as transitional between the ‘‘west coast’’ climatic zone and the ‘‘western hills and mountains’’ variant of the ‘‘central uplands’’ climatic zone to the east (Banfield, 1983). Based on 1991–1997 data from a data collecting platform southeast of the study site (488230 N, 578340 W, 420 m asl), mean annual temperatures are 2.3 8C, with total precipitation of 1377 mm, of which 29% is snow (K. Rollings, personal communication). The study site lies within the generally very rugged, productive Corner Brook boreal forest section B28b (Rowe, 1977) and on the western edge of the Corner Brook subregion of the western Newfoundland ecoregion (Damman, 1983; Meades and Moores, 1994). The Damman forest types found throughout the study area include Dryopteris-Lycopodium-balsam fir, Dryopteris-Hylocomium-balsam fir, Hylocomium-balsam fir, Pleurozium-balsam fir, Kalmia-black spruce and black sprucefeathermoss (Damman, 1967; Meades and Moores, 1994). 2.2. Plot selection, sampling and measurement An extensive aerial and ground reconnaissance combined with photo interpretation of 1997 1:12,500 colour aerial photographs of the study area by the Newfoundland and Labrador Forest Service confirmed a landscape-level mosaic of stands ranging in age from age class 1 (0–20 years) to age class 7 (>120 years). Given the general inaccessibility of the study area and the desire to sample stands within a specific range of age variability and stand development, sample plots were chosen arbitrarily without preconceived bias (McCune and Grace, 2002). At least five candidate stands within each interpreted age class were identified on the aerial photos and located on the ground. The interpreted stand ages served only as a basis for sampling and, if need be, were subsequently refined by standlevel tree age analysis. A total of 50 stands were sampled (Fig. 1). Prior to plot layout, a thorough reconnaissance was made of each candidate stand and its surrounding area. Only structurally homogeneous stands (sensu Poore, 1962) with no history of logging were sampled. Care was taken to minimize overlap of different aged stands. The first corner of each sample plot ranging from 100 to 750 m2 was randomly located. All live and dead trees 1.3 m tall were flagged, numbered, identified to species and measured for caliper diameter (cm) at breast height (dbh) and mortality status. Tree height (0.1 m) was obtained by measuring the length of dominant or codominant trees felled for stem and dendrochronological analysis. 2.3. Self-thinning regression analysis Empirical analysis and verification of the 3/2 self-thinning rule depends in particular on three components of the statistical

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

297

cohort-bimodal stands, the four classic stages of stand development were identified: (1) stand initiation, (2) stem exclusion, (3) stand re-initiation and (4) transition-true oldgrowth. Single-cohort and bimodal stands formed a welldefined 100–120-year chronosequence. Bistaged stands included those lower site quality stands characterized by a regenerating pre-self-thinning stratum overtopped by large black spruce veterans. Reverse-J stands were generally allaged stands created by periodic insect-caused partial canopy mortality and subsequent gap or patch regeneration, or they were simply low site, edaphic reverse-J stands dominated by layered black spruce and slow-growing balsam fir. Caution is important in the use of static stand measurements to infer stand dynamics. While the use of chronosequences promises special advantages in the study of long-term stand dynamics, care must be taken to minimize the confounding effects of site heterogeneity and disturbance history (Cole and van Miegroet, 1989; Pickett, 1989). The use of chronosequences assumes that all stands were of similar condition at the same age in terms of site type and quality, disturbance history and species composition. To confirm that the single-cohort stands formed a natural chronosequence, site indices were calculated for each sample stand using methodologies provided by Page (1968), Newton (1992) and Pothier and Savard (1998). No significant differences existed among the three methods used to calculate site index (P = 0.1696, n = 132, two-way ANOVA with Tukey-Kramer adjusted P values for unbalanced data). The range of site indices for the selfthinning stands (Table 1) was such that these stands were deemed a legitimate chronosequence. Table 2 provides a detailed mensurational description of the chronosequence plots by age class. Successful parameter estimation of a linear log10 density–log10 mean tree volume relationship is possible only when actively self-thinning stands are examined (Weller, 1987a; Lonsdale, 1990). Given that no well accepted a priori method of stand selection exists (Bi et al., 2000), only single-cohort stands having a Weibull distribution shape parameter between 1.5 and 3.6 (Bailey and Dell, 1973) and a kurtosis value greater than 1.2

Fig. 1. Location of sample plots. Non-forested areas include peatlands and heathlands, as well as scrub forests not capable of producing 30 m3 gross merchantable volume ha1 at rotation age.

methodology used: (1) the selection of the self-thinning database, (2) the mathematical formulation of the self-thinning rule and (3) the regression algorithm used to fit the thinning line (Weller, 1987a; Sackville Hamilton et al., 1995). These three components were considered in the following manner: (1) Selection of self-thinning stands: Four main stand structural types were identified across the landscape: (1) single-cohort stands, (2) bimodal stands, (3) bistaged stands and (4) reverse-J stands (McCarthy, 2004). Among the singleTable 1 Summary of mensurational characteristics of self-thinning stands Plot

Average stand age (years)

Site index a

Live (trees ha1)

Skew

Dead (trees ha1)

Skew

19 9 2 4 43 10 20 25 3 16 13 29

60.5 64.8 71.1 72.8 75.9 79.5 81.7 84.2 87.1 89.8 92.5 109.2

9.3 8.5 9.1 10.6 9.9 10.2 10.2 13.0 11.4 10.6 11.2 11.4

31,500 30,667 15,467 7,333 10,225 10,425 4,444 3,050 6,311 7,700 6,350 4,200

1.56 1.97 0.82 0.62 1.30 1.07 0.08 0.11 0.68 0.95 0.83 0.68

17,200 4,133 10,089 10,089 9,950 9,500 9,600 4,575 6,044 7,475 9,125 3,250

3.58 3.05 3.98 3.76 4.09 3.66 2.01 2.17 3.56 4.10 2.38 2.75

a

Height of dominant trees at breast height age of 50 years. Site index values are derived from Page’s (1968) site index curves developed for western Newfoundland.

c

S.D. Mean

A, Hd, Dq, G, Vt and N refer to total stand age, dominant tree height, quadratic mean diameter, basal area, total stand volume and live tree density. S.D. denotes standard deviation. Total stand volumes estimated using the balsam fir equation given for Forest Management Unit 15 in western Newfoundland (Warren and Meades (1986)). a

b

A (year) Hd (m) Dq (cm) G (m2 ha1) Vt (m3 ha1)c N (stems ha1)

Mean Mean Mean

S.D.

Mean

S.D.

S.D. Mean

S.D.

Mean

S.D.

Mean

S.D.

Mean

S.D.

111–120 101–110 91–100 81–90 71–80 61–70 31–40 21–30

b

11–20

Variablea (units) Stand age class (years)

Table 2 Summary of the mensurational characteristics of sample plots by age class from the 120-year chronosequence

17 6 24 7 31 11 63 7 75 9 86 8 97 8 106 9 114 13 2.89 0.44 4.25 1.49 3.40 0.81 7.11 0.92 9.41 1.45 12.08 1.53 14.67 1.74 17.02 1.78 17.95 2.35 2.34 1.44 3.66 1.45 7.51 17.76 4.32 1.00 7.75 1.82 10.99 4.50 15.39 6.65 21.64 29.91 21.85 26.03 4.81 3.93 11.12 5.49 17.14 – 47.65 8.34 50.00 2.45 51.53 6.11 43.15 7.10 42.98 4.24 40.06 0.48 7.62 6.92 24.69 12.00 29.99 – 166.05 42.26 222.40 33.50 283.82 47.60 278.46 41.89 312.73 23.91 302.38 38.18 10,975 8,217 14,032 13,861 3,875 – 31,083 589 10,862 3,379 5,376 2,046 2,448 1,666 1,119 204 1,110 348

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309 S.D.

298

(Wyszomirski, 1992) were retained (Be´gin et al., 2001). The self-thinning character of the sample stands was confirmed by an assessment of selected stand mensurational characteristics, including skewness of the diameter distributions for both live and dead trees, as well as relative live and dead tree densities (Table 1). The size–density relationship of the old bimodal stands as well as a group of post-self-thinning modal stands were also of interest. The old modal stands were identified as those stands showing a noticeable decline in total live basal area with evidence of stand break-up by windthrow and stem breakage. This occurred in stands having reached a height of approximately 15 m and a total age of approximately 90 years. The size–density relationship of these older stands was assessed independently of the size–density relationship of the stands judged to be actively self-thinning. (2) Mathematical formulation of the self-thinning rule: The self-thinning rule was expressed in terms of mean tree volume (or biomass) as w ¼ kd 3=2 (where w ¼ mean plant mass, k = a parameter, d = plant density and 3/2 = a constant independent of plant species). Balsam fir tree volumes (m3) were estimated using the equation provided for Forest Management Unit 15 in western Newfoundland (Warren and Meades (1986)). (3) Regression algorithm: Of critical importance to the debate about the universality of the self-thinning rule has been the comparison of the empirically calculated regression slope with the theoretical slope of 1.5 first suggested by Yoda et al. (1963). Different regression algorithms will produce different slope estimates, depending on the error structure of the data and the correlation between the bivariate data (Sackville Hamilton et al., 1995; Bi, 2004; Zhang et al., 2005). All self-thinning bivariate relationships were analyzed with Model II simple linear regression using reduced major axis (RMA) and ordinary least squares (OLS) (Legendre, 2001). Permutation tests using 999 random permutations were used to test for significance except in the case of RMA, which cannot be formally tested. In practice, this poses no problem given that the null hypothesis that the population correlation coefficient equals zero is essentially the same (McArdle, 1988; Quinn and Keough, 2002). For reduced major axis, the confidence interval of slope followed Jolicoeur and Mosimann (1968) and McArdle (1988). Each of the log10–log10 linear regressions of mean tree volume, quadratic mean diameter (qmd) and tree height versus tree density were compared for two groupings: (1) all singlecohort stands 60–90 years old and (2) all single-cohort stands > 90 years and all bimodal stands. Stands 60–90 years of age were judged to be undergoing active self-thinning, whereas the older stands (>90 years) were judged increasingly influenced by density-independent mortality processes. The log10 qmd–log10 density relationships of the two stand types were also compared. For each regression couplet of interest, heterogeneity of slopes was assessed by a covariance analysis of both the treatment and the X covariate (tree density) (Littell et al., 2002)

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

299

after fitting by ordinary least squares. A significant interaction of treatment with tree density confirmed heterogeneity of slopes and a separation of the stands into two distinct groupings (60– 90 years and >90 years including bimodal). Normality and variance assumptions were tested by standard residual and statistical analysis.

spruce stands in western Newfoundland. The SDMD was graphically assessed for conformance to self-thinning behavior and general stand development phases as inferred from temporary chronosequence plot measurements.

2.4. Basal area trends

3.1. Size–density relationships

A locally weighted regression (loess) was used to smooth bivariate basal area trends for both live and dead trees in a chronosequence of single-cohort and bimodal stands (Trexler and Travis, 1993). Choice of the proper smoothing parameter was determined by plotting Y residuals against X and applying loess smoothing to the residuals. Based on loess analysis of the residual plots, a span of 0.5 was chosen as the value that provided a smoothing curve that best reflected the pattern in the data. Loess smoothing was performed using Sigmaplot 8.0 (SPSS Inc., 2002), which uses a tri-cube weighting function as the default.

Self-thinning dynamics of the chronosequence of natural stands are presented in both the classic linear log10 stand attribute–log10 stem density regression format, as well as within the context of a balsam fir-spruce SDMD developed for stands in the study area. Attributes considered were mean tree volume, as well as live and dead quadratic mean diameter. Single-cohort stands < 60 years of age and with 75 percentile heights of 2.9–6.5 m were located outside the zone of imminent competition mortality (ZICM) and had yet to undergo any appreciable degree of self-thinning. Stands older than approximately 90 years of age and with 75 percentile heights of 14.7–20.2 m were judged to be, more or less, beyond the stage of active self-thinning and increasingly influenced by density-independent mortality processes. Consistent with the well-known principle of self-thinning, live quadratic mean diameter, mean tree volume, as well as

2.5. Stand density management diagram Mean tree volume–stand density data from all single-cohort, bimodal and bistaged stands were plotted unto the SDMD developed by Sturtevant et al. (1998) for mixed balsam fir-black

3. Results

Fig. 2. Log10–log10 mean tree volume/size–density relationships for a chronosequence of balsam fir stands. For (A) mean tree volume and (B) live quadratic mean diameter, distinct regressions are given for stands aged 60–90 years and for stands > 90 years (including old bimodal stands). For (C) dead quadratic mean diameter only one equation is given for all stands > 60 years of age. Regression parameters, 95% confidence intervals and fit statistics are given in Table 3.

300

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

Table 3 Regression parameters, 95% confidence intervals (CI) and fit statistics for log10–log10 mean tree volume/size–tree density self-thinning models Variable

na

Modelb

Intercept (95% CI)

Slope (95% CI)

R2

Mean tree volume (60–90 years) (rc = 0.9886)

12

RMA OLS

6.479 (5.970, 7.045) 6.421 (5.882, 6.960)

1.282 (1.426, 1.153) 1.267 (1.404, 1.131)

0.9772

Mean tree volume (90 years-bimodal) (r = 0.9276)

16

RMA OLS

5.594 (4.961, 6.378) 5.355 (4.645, 6.064)

1.045 (1.293, 0.845) 0.967 (1.193, 0.746)

0.8604

Mean live quadratic mean diameter (60–90 years) (r = 0.9914)

12

RMA OLS

2.932 (2.756, 3.126) 2.915 (2.730, 3.100)

0.509 (0.558, 0.464) 0.504 (0.551, 0.457)

0.9829

Mean live quadratic mean diameter (90 years-bimodal) (r = 0.9137)

16

RMA OLS

2.582 (2.312, 2.921) 2.469 (2.164, 2.774)

0.413 (0.520, 0.328) 0.377 (0.473, 0.281)

0.8348

Mean live quadratic mean diameter (60 years-bimodal) (r = 0.9914)

28

RMA OLS

2.731 (2.649, 2.818) 2.717 (2.632, 2.803)

0.459 (0.484, 0.435) 0.455 (0.479, 0.431)

0.9830

Mean dead quadratic mean diameter (60 years-bimodal) (r = 0.9097)

28

RMA OLS

2.670 (2.399, 2.990) 2.510 (2.212, 2.808)

0.505 (0.596, 0.427) 0.459 (0.544, 0.375)

0.8275

All parameters are given in their log10 form. All regressions are significant at P < 0.0001. a n: Number of stands. b RMA: Reduced major axis, OLS: ordinary least squares. All regressions carried out using the Model II regression software provided by Legendre (2001). c r: Pearson product-moment correlation coefficient.

dead quadratic mean diameter displayed significant log10–log10 linear relationships with tree density (Fig. 2; Table 3). These models were constructed using data from both modal stands undergoing active, intense self-thinning as well as modal and bimodal stands experiencing increasing levels of densityindependent mortality. For mean tree volume, the ordinary least squares slope of 1.267 for the youngest range of stands was significantly less than the slope of 0.967 determined for the oldest stands (P = 0.0221). Two distinct regression lines were also identified for log10 live quadratic mean diameter, with the youngest and oldest stands having significantly different ordinary least squares slopes of 0.504 and 0.377, respectively (P = 0.0142). Regression slopes and intercepts of the dead quadratic mean diameter relationships were not statistically different (P > 0.05) for the younger and older stands; therefore, only one equation was presented. For both tree volume and live quadratic mean diameter, distinct reduced major axis regressions were recognized for self-thinning and non-self-thinning stands. For the selfthinning tree volume–density line, the slope of 1.282 was significantly different from the theoretical self-thinning slope of 1.5 (P < 0.05). Older stands judged to be growing beyond self-thinning had a shallower slope of 1.045. The reduction in slope for the older stands indicated a reduction in mortality rate compared to the actively self-thinning younger stands. The live quadratic mean diameter–density relationship was similar, with the self-thinning slope of 0.509 greater than the slope of 0.413 for the older stands. The slope of the dead quadratic mean diameter regression for all stands combined was 0.505. Of note is the comparison between the live and dead quadratic mean diameter as a function of stand density for all stands (Fig. 3). Covariance analysis of parameters estimated by OLS analysis showed the regression lines to be parallel (P = 0.9193) with significantly different intercepts (P < 0.0001). For live

trees, RMA intercept and slope were calculated as 2.731 (95% CI = 2.649, 2.818) and 0.459 (95% CI = 0.484, 0.435), respectively. For dead trees, regression slope and intercept were calculated as 2.670 (95% CI = 2.399, 2.990) and 0.505 (CI = 0.596, 0.427). Across the range of stand densities examined, the live trees had consistently larger diameters than the dead trees. In other words, smaller trees had a greater mortality rate than larger trees. 3.2. Stand density management diagram Stand development among all chronosequence stands is best considered within the context of a SDMD developed for mixed balsam fir-spruce stands in western Newfoundland (Sturtevant et al., 1998) (Fig. 4). Four types of stand development are proposed:

Fig. 3. Reduced major axis regression for the log10–log10 live and dead quadratic mean diameter–tree density relationship for a chronosequence of balsam fir stands (60 years-bimodal). Regression parameters, 95% confidence intervals and fit statistics are given in Table 3.

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

301

3.3. Basal area trends Basal area trends across the chronosequence were assessed in terms of both 75 percentile stand height and total stand age. For each independent variable examined, the relationships with both stand height and stand age were similar. For both live and dead basal area trends, three distinct phases in stand development were identified (Fig. 5A):

Fig. 4. Relative position of 41 sample stands on the mixed balsam fir-black spruce stand density management diagram given by Sturtevant et al. (1998). Diagram based on a relative density index (Pr) = 1.0, a lower limit of the zone of imminent competition mortality (ZICM) at Pr = 0.5 and an approximate crown closure (CC) line at Pr = 0.13. Isolines of both tree height and quadratic mean diameter are also indicated. Relative density index is defined as the ratio of actual stand density to the maximum stand density attainable in a stand with the same mean tree volume (Drew and Flewelling, 1979) and is independent of site quality.

(1) Single-cohort and bimodal stands undergoing intense and active self-thinning (9, 19, 2, 10, 43, 4, 16, 13, 3, 20, 29 and 25). All stands undergoing active self-thinning were located relatively close to the upper limit of the ZICM and paralleled, more or less, the maximum size–density line at a slope of 1.28. Most of the stands were 60–90 years old, all had 75 percentile height values ranging from 7.1 to 14.4 m and stem densities > 3000 stems ha1. (2) Single-cohort and bimodal stands past intense, active selfthinning and beginning or undergoing stand break-up (22, 32, 40, 27, 50, 31, 34, 21, 11, 12, 37, 36, 35, 48, 24 and 7). These stands were more or less located close to or outside the lower limit of the ZICM. All stands were >90 years of age, had stem densities < 3000 stems ha1 and tracked along at a significantly reduced slope of 1.04. (3) Single-cohort stands showing no evidence of active selfthinning (8, 6, 51, 38, 39, 42, 23, 15 and 46). All nine stands were <40 years of age and ranged in 75 percentile height from 2.9 to 6.5 m. All stands were below the lower limit of the ZICM. Except for plot 39 with the highest density of 39,000 stems ha1, all other stands were either close to, or even below, the crown closure line. (4) Bistaged stands (49, 45, 26 and 44). Except for plot 44, these stands were either on or below the lower limit of the ZICM. Because of low stocking levels that were age and site-related (low site index), these stands were not undergoing active self-thinning.

(1) An aggrading phase characterized by rapid accumulation of live basal area to approximately 9–10 m stand height (60 years total age). Live basal area ranged from a minimum of 1 m2 ha1 in the youngest stands of age class 11–20 years to maximum values exceeding 50 m2 ha1. Changes in total dead basal area followed an opposite trend with decreases from a peak of 23 m2 ha1 in the youngest stands to minimum levels near 10 m2 ha1 in stands around 60 years of age. (2) A transition period of slightly declining, but relatively stable, live basal area levels from 10 to 15 m (60–90 years of age). During this period of stand development, dead basal area accumulated at a steady rate to levels approximately equal to those of the early post-disturbance period. (3) A degrading phase characterized by an increased rate of live basal area decline for stands 15–20 m in height and 90–140 years of age. Levels of dead basal area followed a similar decline to that of live basal area, but for stands approximately 2 m taller and 10–20 years older. Total basal area trends were mainly the result of changes in the basal area of balsam fir. Basal area levels of black spruce, white spruce (Picea glauca (Moench) Voss) and white birch (Betula papyrifera Marsh) showed no appreciable trends across the chronosequence. Live and dead basal area values of these species rarely exceeded 10 and 5 m2 ha1, respectively. The relative proportions of total live and dead basal area showed opposing linear trends across the chronosequence (Fig. 5B). The relative proportion of total dead basal area declined steadily from a high of approximately 80% in the youngest stands to a low of approximately 20% in 60-year-old stands and then increased to levels not exceeding an average of 30% in the oldest and tallest stands. The trend was reversed for live total basal area. The relative proportion of total live basal area increased from approximately 20% in the youngest stands, peaked at around 80% in the 60-year-old stands and then decreased to average levels not <70% in the oldest and tallest stands. For bimodal, reverse-J and bistaged stands, the proportions of the total dead basal area were almost equal at 23.3%, 24.3% and 23.0%, respectively. 3.4. Stand development Fig. 6 shows the changes in tree diameter frequency by stand age class across the established chronosequence of 31 singlecohort stands and 6 bimodal stands. Evident was the shift in dbh frequency distributions from a reverse-J distribution in the

302

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

Fig. 5. Stand basal area as a function of stand 75 percentile height and total stand age for all single-cohort and bimodal stands. (A) Total live and dead basal area and (B) relative proportion of total live and dead basal area.

youngest stands to modal stands that exhibited increasing standard deviation and flattening of the normal distribution as average tree density decreased and both average tree diameter and diameter range increased. In the old bimodal stands, the development of a distributional bimodality highlighted the recruitment of smaller diameter trees within canopy gaps as the stands began to break up. 4. Discussion 4.1. Self-thinning rule Increased inter-tree competition due to crown closure and light extinction produced classic changes in tree plasticity and mortality as defined by the 3/2 power law (Yoda et al., 1963; White and Harper, 1970; Westoby, 1984). Table 4 compares the self-thinning slope and intercept found in this study with those determined for other self-thinning Abies stands. The slope of 1.282 differed significantly from the assumed selfthinning slope of 1.5, comparing favourably to that

determined by Mohler et al. (1978), Sprugel (1984) and Wilson et al. (1999). Recent empirical and theoretical research on allometric scaling and mass properties in plants affirmed that the self-thinning slope should be 4/3 and not 3/2 (Lonsdale, 1990; Enquist et al., 1998; Franco and Kelly, 1998), thus making the study slope of 1.282 of greater theoretical acceptability. A number of other forest tree self-thinning studies have also found slope coefficients significantly less than the theoretical 1.5 (Osawa and Allen, 1993; Kenkel et al., 1997). Early formulations of the self-thinning rule focused on the seeming law-like inviolability and universality of the 3/2 asymptotic self-thinning slope (White, 1980). Subsequent research, however, has shown that the self-thinning exponents can be more variable than originally assumed, and that the exponent value may differ from the theoretical value depending on species, sample size, volume equations, regression algorithm, site quality, plant allometric relationships and whether the self-thinning data set actually includes stands representing the maximum combination of size and density (Zeide, 1987,

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

303

1991; Weller, 1987a,b, 1989; Sackville Hamilton et al., 1995; Bi et al., 2000). For example, the tree volume equation used in different studies may have profound effects on the log10(mean tree volume)–log10(tree density) relationship. The slope coefficient was 1.41 in Solomon and Zhang (2002), 1.21 in Wilson (1996) and 1.282 in this study. However, if one compares the slope coefficients for the log10(quadratic mean diameter)–log10(tree density) relationships, one will find 0.52 for Solomon and Zhang (2002), 0.51 for Wilson (1996) and 0.509 for this study. 4.2. Self-thinning, tree mortality and size distribution The good fit of the empirical chronosequence data with the mixed balsam fir–black spruce SDMD produced for western Newfoundland (Sturtevant et al., 1998) provides the first independent test of the general validity of the SDMD for insectdriven balsam fir forests of western Newfoundland. The SDMD as a good reflection of stand dynamics in balsam fir-black spruce stands, and may therefore act as an effective operational tool for control of stand density in forest management. Selfthinning dynamics of the study stands are assessed with the help of this SDMD. Three distinct phases in tree mortality were evident across the chronosequence: (1) post-insect disturbance, (2) active selfthinning and density-dependent tree mortality and (3) post-selfthinning and density-independent mortality. The combined action of these three processes is often best described by a sigmoidal relationship, a relationship that was clearly evident during post-disturbance stand development of the balsam fir stands. U-shaped patterns of coarse woody debris accumulation have been described for a number of forests (Spies et al., 1988; Clark et al., 1998; He´ly et al., 2000).

Fig. 6. Stand structural development across a chronosequence of single-cohort and bimodal stands. The number in brackets after the age class denotes the number of stands per age class.

4.2.1. Post-insect disturbance The young single-cohort stands (<40 years old) all lie outside the ZICM, with two stands located below the crown closure line. Prior to crown closure, the site is not fully occupied, competition among the trees is limited and mortality is minimal or non-existent, with tree growth essentially independent of stand density. A distinct hierarchy of dominance and suppression had not yet developed. Following crown closure, trees begin to compete with each other, with onset of self-pruning and crown class differentiation. Dead volume was dominated by snag legacies still standing after historic insect outbreaks, and comprised up to 80% of the total live and dead tree basal area. Pre-self-thinning distributions are generally truncated reverse-J structures, with the L-shaped distribution characteristic of the youngest age class of saplings (Fig. 6). The truncated reverse-J shaped diameter distributions may simply be an expression of the exponential growth of saplings growing free of competition (Koyama and Kira, 1956). Even before the onset of self-thinning, there is some degree of tree size differentiation due to local neighbourhood effects including microsite effects, differential emergence effects and genetic differences that establish, right from the beginning, a dominance hierarchy that

304

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

Table 4 Comparison of Abies self-thinning in the Northeast United States, eastern Canada and Japan Location

Stand history

Species

Mount Shimigare, Central Japan

Wave forest

Abies veitchii (Veitch’s silver fir) Abies mariesii (Maries fir)

Whiteface Mountain, New York

Wave forest

Abies balsamea (Balsam fir)

Mount Shimigare, Central Japan

Wave forest

Whiteface Mountain, New York

Dependent variable

Slope

Intercept Regression Self-thinning algorithm line

Aboveground dry weight (kg)

1.45

–

OLSa

Dynamic thinning line

Kuroiwa (1959)

23 Mean weight (g) Trunk only (g)

1.22 1.30

– –

PCAb

–

Mohler et al. (1978)

Abies veitchii Abies mariesii

14 Mean D0.1c 2Hd

1.45

6.42

OLS

Dynamic thinning line

Kohyama and Fujita (1981)

Wave forest

Abies balsamea

29 Mean tree mass (g) 1.24 Bole mass (g) 1.43

3.94 3.65

OLS

Dynamic thinning line

Sprugel (1984)

Northern Newfoundland Wave forest

Abies balsamea

2.7

–

Northern Maine

–

Dynamic thinning line Maximum boundary line

Robertson (1993) Wilson et al. (1999)

Eastern Quebec

Cutovers

Picea rubra (red spruce)Abies balsamea Abies balsamea

12 Mean tree mass (kg tree1) 21 Mean stem volume (ft3)

Northeast USA

Climax forests

Maximum boundary line Maximum boundary line Dynamic self-thinning

Be´gin et al. (2001) Solomon and Zhang (2002) This study

Western Newfoundland a b c d e

Picea rubra-Abies balsamea Insect-mediated Abies balsamea chronosequence

n

801 Mean total bole volume (m3) 4559 Mean tree volume (m3) 12 Mean tree volume (m3)

1.5

1.215 4.52

RMAe

1.441 4.114

Robust PCA RMA

1.41

8.44

1.282 6.479

RMA

Reference

OLS: Ordinary least squares. PCA: Principal component analysis. D0.1: Tree diameter (cm) at the height (H) of H/10 from the ground level. H: Tree height (m). D0.1 2H used as a substitute for tree weight (cm2 m). RMA: Reduced major axis.

can influence the future development of each individual (Watkinson et al., 1983; Weiner, 1984). This difference will attenuate after crown closure and the onset of competition mortality. 4.2.2. Active self-thinning Stands located within the ZICM tracked parallel to the upper self-thinning boundary line and underwent active and accelerated self-thinning. The upper limit of the ZICM defines a maximum size–density line at which point developing stands experience a reduction in density and an increase in stem diameter as they track more or less parallel to the maximum size–density line (Long and Smith, 1984; Jack and Long, 1996). Self-thinning occurs when stands have assumed maximum site occupancy, full crown closure and maximum biomass at any particular density. Continued stand growth will occur only if tree mortality occurs. Prior to and during the process of self-thinning, stands undergo significant structural change. First, is the production of dead trees that die during the intense, light-driven, asymmetric competition and secondly, is the dramatic shift in the dbh frequency of live trees. Changes in tree size hierarchy and skewness of diameter distributions with stand development creates the conditions necessary for self-thinning. Self-thinning commenced once stands had reached quadratic mean diameters of 4–5 cm, 75 percentile heights of 7–8 m and stem densities of approximately 31,000 stems ha1. These stand conditions were reached around 60–65 years of age. The

beginning of self-thinning corresponds to maximum stand volume conditions. Yield curves for balsam fir stands (medium site, crown density 50–75%) in the study area have maximum volumes of 140–150 m3 ha1 developing in stands of stump age > 60 years (Boyd Pittman, Newfoundland Forest Service, personal communication). With increase in stand height, the size–density relationships closely followed the 3/2 power self-thinning rule. The trajectory of the stands along the upper boundary thinning line, however, slowed as the stands developed. At a particular section of the self-thinning curve, the stands began to fall away from the self-thinning upper boundary line and to follow a line with slope approximating unity. This decrease in the size–density slope from a self-thinning line of 1.5 to a non-self-thinning line of 1.0 has been observed in other plant size–density relationships (Peet and Christensen, 1980; Hutchings and Budd, 1981; White, 1981; Westoby, 1984) including balsam fir (Sturtevant et al., 1997; Be´gin et al., 2001). As stands move into self-thinning, the skewness of the diameter distribution becomes increasingly positively skewed (Ford, 1975; Mohler et al., 1978; West and Borough, 1983; Newton and Smith, 1988; Xue and Hagihara, 1999). This increase in skewness is caused by the higher mortality rate of the smaller trees, a measure of the increasing competitive advantage of the taller trees. With an increase in tree height and canopy differentiation, the degree of competitive interaction increases until the development of intense interspecific competition during the self-thinning process (Long and Smith,

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

1984; Peet and Christensen, 1987). Inter-specific competition during self-thinning is generally one-sided, asymmetric (resource pre-emption) competition in which individuals share limited resources, particularly light, disproportionate to their relative size (West et al., 1989; Newton, 1990; Weiner, 1990; Newton and Jolliffe, 1998a,b). Many studies have confirmed the greater probability of death for the lower proportion of plant size distributions during self-thinning (Ford, 1975; Yarranton and Yarranton, 1975; Peet and Christensen, 1987; Kikuzawa, 1988; Kenkel et al., 1997; Ogawa and Hagihara, 2003). 4.2.3. Post-self-thinning Self-thinning no longer occurs in stands >90 years old, with 2000–3000 stems ha1 and 15–17 m in height. Along this new shallow slope trajectory, density-independent mortality processes including butt rots and root rots, windthrow and insect herbivory, increase in importance. During self-thinning and a gradient of 1.5, total stand volume increases despite the decrease in stem density. However, with the change in the selfthinning slope to 1, the total yield becomes constant (White and Harper, 1970), with net biomass production approaching or matching mortality loss. Given the noticeable beginnings of decline in basal area at around 90 years and stand heights of 15– 17 m, it seems that the stands have entered the stand break-up stage at this point. After the period of intense self-thinning and during the early period of density-independent mortality, diameter distributions are generally symmetric in character (Mohler et al., 1978; Lorimer and Frelich, 1998). With stand development and increased density-independent mortality, bimodality develops in response to the release of advance regeneration in the created canopy gaps. 4.3. Previous insect damage and self-thinning Given the long history of insect herbivory in the study area, the sampled stands may have incurred density-independent mortality over and above that due to density-dependent mortality caused by self-thinning. This possible reduction in crown closure in self-thinning stands as a result of insect damage could produce a flattening of the thinning line away from a slope of 1.5. Within the study area region, hemlock looper outbreaks occurred during the periods, 1946–1955, 1959–1964, 1966– 1972 and 1983–1995 (Otvos et al., 1979; Hudak et al., 1996). A dendrochronological analysis of historic insect outbreaks in the region close to the study area confirmed the occurrence of decadal-level light to moderate infestations of both spruce budworm and hemlock looper during the 19th and 20th centuries (Jardon and Doyon, 2003). Many of the sampled stands would have been partially (or fully) defoliated during the late 1970s–early 1980s spruce budworm outbreak. Areas of dead, moribund and very seriously damaged stands were mapped in the study area, particularly along the steep forested slopes facing Grand Lake and Little Grand Lake (Hudak and Raske, 1981; Raske et al., 1982).

305

Parts of the study area were sprayed three times against the hemlock looper since 1969. In 1969, small northwestern and southeastern sections of the study area were sprayed with two applications of fenitrothion. Much of the eastern section of the study area was sprayed in 1987 (two applications of fenitrothion) and 1995 (two applications of Bacillus thurengiensis var kurstaki (B.t.)) (Hubert Crummey, Newfoundland Forest Service, personal communication). No spray program was carried out during the spruce budworm outbreak of the late 1970s–early 1980s. Given the lack of data on the stand-level defoliation rates and the impact of past protection measures on tree mortality, it is difficult to state to what degree previous defoliation and consequent canopy disturbance have impacted the derived selfthinning inferences. Suffice it to say that retrospectively differentiating these different sources of tree mortality effects is problematic when using static plot measurements. 4.4. Stand break-up dynamics In western Newfoundland, maximum merchantable volumes usually occur in stands of approximately 60–70-year-old balsam fir and 100–110-year-old black spruce (Fig. 7). Catastrophic stand break-up, at least in terms of gross merchantable volume, is around 130 years for balsam fir and 160–180 years for black spruce. While our study did not examine the break-up of stands, the stand chronosequence examined provides some insight into the later stages of stand development. No evidence was found of catastrophic stand break-up (other than insect kill) as depicted in the modeled yield curves. Data on stand break-up in Newfoundland is sketchy and is the least known stage of stand dynamics (Boyd Pittman, Newfoundland Forest Service, personal communication) Other than catastrophic stand kill by insects, evidence points to a more gradual and prolonged break-up, if indeed that is what actually happens. If old stands are not subject to insect-induced mortality, then stands may eventually shift to transition old-growth into the old-growth stage of stand development with the maintenance of significant crown cover and standing volume. Surprisingly, in the western Newfoundland yield curves, there were equal break-up ages for stands of good, medium and poor site quality. For balsam fir, the curves showed a gradual decline in merchantable volume around 120 years, followed by a drastic drop around 140 years, implying catastrophic stand break-up. In general, the break-up ages did not vary with site quality because this portion of the yield curves was not well supported by empirical data. It was not until 1992 that the Newfoundland Forest Service extended its permanent sample plot system to include mature and overmature stands (Darrell Harris, Newfoundland Forest Service, personal communication). It is well known that stand break-up and tree longevity is site-dependent, with stands growing on the better sites breaking up earlier than stands growing on the poorer sites (Robichaud and Methven, 1993). This propensity to faster break-up is often attributed to the relationship between tree height and stability

306

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

Fig. 7. Yield curves (gross merchantable volume m3 ha1) for balsam fir and black spruce, western Newfoundland. All curves are for stand density class 2 (51–75% crown closure) and for good, medium and poor site quality stands. Permission to use granted by the Newfoundland and Labrador Forest Service.

(Smith et al., 1987). All poor study sites had reverse-J diameter distributions, complex canopy structure and lower tree height– age ratios compared with stands growing on better sites that had modal diameter distributions, relatively homogeneous canopy structure and higher height–age ratios. These relative stand characteristics would give the poorer site quality stands greater mechanical stability, therefore promoting greater stand longevity. If decline in basal area may be used as a proxy measure of the beginnings of stand break-up (volume loss), then results of this study agree well with established understanding of balsam fir stand dynamics in Newfoundland. Gross merchantable volume normally declines around 70 years of age (measured at breast height), which is close to that documented in the present study that showed declines in basal area at 90 years of age (total age). Stand decline or decreasing volume production occurred at 70 years for balsam fir-spruce stands (Taylor and MacLean, 2005). Natural stand decline beginning at 70–90 years have also been reported by Bakuzis and Hansen (1965) and Porter et al. (2001). These periods are equivalent given the roughly 20 years needed for balsam fir to reach breast height in these stands. The onset of stand break-up for forests of northern Newfoundland was determined to be over 100 years (Page et al., 1970). Stands in western and eastern Newfoundland exhibit maximum mean annual increments at approximately 50 to 65 years of age (Bajzak et al., 1968; Page et al., 1970). 5. Conclusion The balsam fir forests of western Newfoundland are disturbed on a regular basis by outbreaks of the hemlock looper and the spruce budworm. In the Little Grand Lake Provisional Ecological Reserve, recurrent outbreaks over the past century and more have created a chronosequence of stands covering a period of at least 120 years. The chronosequence provided a unique opportunity to examine balsam fir stand

dynamics over this period. The balsam fir-dominated stands followed the classic pattern of self-thinning as defined by the 3/2 power law. A self-thinning slope of 1.28 was significantly different from the theoretical slope of 1.5 and compared well with other studies in Abies-dominated stands. Stands past the stage of self-thinning had a tree volume-tree density slope approaching unity. Balsam fir began to self-thin in stands approaching a 75 percentile height of  7 m and 60 years of age. Self-thinning ceased in 90-year-old stands with 75 percentile heights approaching 15 m. Results of this study confirmed the utility of a mixed balsam fir-black spruce SDMD designed for forests of western Newfoundland (Sturtevant et al., 1998). Acknowledgements This study was supported by the Natural Sciences and Engineering Research Council of Canada, Newfoundland Forest Service, Canadian Forest Service, Western Newfoundland Model Forest, Parks Canada, Corner Brook Pulp and Paper Limited and the Upper Canada Province of the Society of Jesus. Scott Taylor of Gros Morne National Park prepared Fig. 1. Field and lab work was ably carried out by Paul Tipple, Justin Basha, Paul Sinyard and Chris Curnew. Christine Ducharme provided professional secretarial support. Thanks to Boyd Pittman of the Newfoundland Forest Service for provision of yield curve data. The comments of Nancy Britton, Peter Newton and two anonymous reviewers significantly improved the manuscript. References Bailey, R.L., Dell, T.R., 1973. Quantifying diameter distributions with the Weibull function. For. Sci. 19, 97–104. Bajzak, D., Bouzane, J.P., Page, G., 1968. A study of the mensurational characteristics of some important forest types of western Newfoundland. Dept. For. Rural Dev., For. Br., For. Res. Lab. Inf. Rep. N-X-7.

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309 Bakuzis, E.V., Hansen, H.L., 1965. Balsam Fir: A Monographic Review. University of Minnesota Press, Minneapolis. Banfield, C.E., 1983. Climate. In: South, G.R. (Ed.), Biogeography and Ecology of the Island of Newfoundland. Monographiae Biologicae, vol. 48. Dr W. Junk Publishers, The Hague, pp. 37–106. Baskerville, G.L., 1975. Spruce budworm: super silviculturist. For. Chron. 51, 138–140. Be´gin, E., Be´gin, J., Be´langer, L., Rivest, L.-P., Trembley, S., 2001. Balsam fir self-thinning relationship and its constancy among different ecological regions. Can. J. For. Res. 31, 950–959. Bergeron, Y., Gauthier, S., Kafka, V., Lefort, P., Lesieur, D., 2001. Natural fire frequency for the eastern Canadian boreal forest: consequences for sustainable forestry. Can. J. For. Res. 31, 384–391. Bi, H., 2004. Stochastic frontier analysis of a classic self-thinning experiment. Austr. Ecol. 29, 408–417. Bi, H., Wan, G., Turvey, N.D., 2000. Estimating the self-thinning boundary line as a density-dependent stochastic biomass frontier. Ecology 81, 1477– 1483. Blais, J.R., 1983. Trends in the frequency, extent, and severity of spruce budworm outbreaks in eastern Canada. Can. J. For. Res. 13, 539–547. Clark, D.F., Kneeshaw, D.D., Burton, P.J., Antos, J.A., 1998. Coarse woody debris in sub-boreal spruce forests of west-central British Columbia. Can. J. For. Res. 28, 284–290. Cole, D.W., van Miegroet, H., 1989. Chronosequences: a technique to assess ecosystem dynamics. In: Dyck, W.J., Mees, C.A. (Eds.), Proceedings IEA/ BE A3 Workshop: Research Strategies for Long-term Site Productivity, August 1988, Seattle, Washington. IEA/BE A3 Report No. 8, For. Res. Inst., New Zealand, Bull., vol. 152, pp. 5–23. Damman, A.W.H., 1967. The forest vegetation of western Newfoundland and site degradation associated with vegetation change. Dissertation. University of Michigan. Damman, A.W.H., 1983. An ecological subdivision of the Island of Newfoundland. In: South, G.R. (Ed.), Biogeography and Ecology of the Island of Newfoundland. Monographiae Biologicae, vol. 48. Dr. W. Junk Publishers, The Hague, pp. 163–206. Drew, T.J., Flewelling, J.W., 1979. Stand density management: an alternative approach and its application to Douglas-fir plantations. For. Sci. 25, 518– 532. Engelmark, O., 1999. Boreal forest disturbances. In: Walker, L.R. (Ed.), Ecosystems of Disturbed Ground. Ecosystems of the World, vol. 16. Elsevier, Amsterdam, pp. 161–186. Enquist, B.J., Brown, J.H., West, G.B., 1998. Allometric scaling of plant energetics and population density. Nature 395, 163–165. Ford, E.D., 1975. Competition and stand structure in some even-aged plant monocultures. J. Ecol. 63, 311–333. Franco, M., Kelly, C.K., 1998. The interspecific mass–density relationship and plant geometry. Proc. Natl. Acad. Sci. 95, 7830–7835. Greene, D.F., Zasada, J.C., Sirois, L., Kneeshaw, D., Morin, H., Charron, I., Simard, M.-J., 1999. A review of the regeneration dynamics of North American boreal forest tree species. Can. J. For. Res. 29, 824–839. He´ly, C., Bergeron, Y., Flannigan, M.D., 2000. Coarse woody debris in the southeastern Canadian boreal forest: composition and load variations in relation to stand development. Can. J. For. Res. 30, 674–687. Holling, C.S., 1992. The role of forest insects in structuring the boreal landscape. In: Shugart, H.H., Leemans, R., Bonan, G.B. (Eds.), A Systems Analysis of the Global Boreal Forest. Cambridge University Press, Cambridge, pp. 170–191. Hudak, J., Raske, A.G. (Eds.), 1981. Review of the spruce budworm outbreak in Newfoundland—its control and forest management implications. Environ. Can. Can. For. Serv. Newfoundland For. Res. Cent. Inf. Rep. N-X-205. Hudak, J., Sutton, W.J., Stone, D.M., O’Brien, D.S., Osmond, S.M., Crummey, H.R., 1996. Outbreak history of eastern hemlock looper in Newfoundland, 1983–1995. In: Hudak, J. (Ed.), The eastern hemlock looper, Lambdina fiscellaria fiscellaria (Guen.) (Lepidoptera: Geometridae) in Newfoundland, 1983–1995. Nat. Resourc. Can., Can. For. Serv., Newfoundland and Labrador Reg. Inf. Rep. N-X-302, pp. 2–6. Hutchings, M.J., Budd, C.S.J., 1981. Plant competition and its course through time. Bioscience 31, 640–645.

307

Jack, S.B., Long, J.N., 1996. Linkages between silviculture and ecology: an analysis of density management diagrams. For. Ecol. Manage. 86, 205–220. Jardon, Y., Doyon, F., 2003. Balsam Fir Stand Dynamics After Insect Outbreak Disturbances in Western Newfoundland Ecoregion (Corner Brook Subregion). Western Newfoundland Model Forest, Inc., Corner Brook, Newfoundland. Johnson, E.A., 1995. Fire and Vegetation Dynamics: Studies from the North American Boreal Forest. Cambridge Studies in Ecology, reprint ed. Cambridge University Press, Cambridge. Jolicoeur, P., Mosimann, J.E., 1968. Intervalles de confiance pour la pente de l’axe majeur d’une distribution normal bidimensionnelle. Biome´trie-Praxime´trie 9, 121–140. Kenkel, N.C., Hendrie, M.L., Bella, I.E., 1997. A long-term study of Pinus banksiana population dynamics. J. Veg. Sci. 8, 241–254. Kikuzawa, K., 1988. Intraspecific competition in a natural stand of Betula ermanii. Ann. Bot. 61, 727–734. Kohyama, T., Fujita, N., 1981. Studies on the Abies population of Mt. Shimagare. I. Survivorship Curve. Bot. Mag. (Tokyo) 94, 55–68. Kohyama, T., Hara, T., Tadaki, Y., 1990. Patterns of trunk diameter, tree height and crown depth in crowded Abies stands. Ann. Bot. 65, 567–574. Koyama, H., Kira, T., 1956. Intraspecific competition among higher plants. VIII. Frequency distribution of individual plant weight as affected by the interaction between plants. J. Inst. Polytech. Osaka City Univ. Ser. D. Biol. 7, 73–94. Kuroiwa, S., 1959. Ecological and physiological studies on the vegetation of Mt. Shimagare III. Intraspecific competition and structural development of the Abies forest. Bot. Mag. (Tokyo) 72, 413–421. Legendre, P., 2001. Model II regression—User’s guide. De´partement de sciences biologiques, Universite´ de Montre´al, p. 23. Available from http://www.bio.umontreal.ca/legendre/indexEnglish.html. Littell, R.C., Stroup, W.W., Freund, R.J., 2002. SAS1 for Linear Models, fourth ed. SAS Institute Inc., Cary, NC. Long, J.N., Smith, F.W., 1984. Relation between size and density in developing stands: a description and possible mechanisms. For. Ecol. Manage. 7, 191– 206. Lonsdale, W.M., 1990. The self-thinning rule: dead or alive? Ecology 71, 1373–1388. Lorimer, C.G., Frelich, L.E., 1998. A structural alternative to chronosequence analysis for uneven-aged northern hardwood forests. J. Sust. For. 6, 347– 366. MacLean, D.A., 1984. Effects of spruce budworm outbreaks on the productivity and stability of balsam fir forests. For. Chron. 60, 273–279. MacLean, D.A., Ebert, P., 1999. The impact of hemlock looper (Lambdina fiscellaria fiscellaria (Guen.)) on balsam fir and spruce in New Brunswick, Canada. For. Ecol. Manage. 120, 77–87. McArdle, B.H., 1988. The structural relationship: regression in biology. Can. J. Zool. 66, 2329–2339. McCarthy, J.W., 2004. Natural disturbance and structure in two primary boreal forests of western Newfoundland. Department of Forest Sciences, Faculty of Forestry. Dissertation. University of British Columbia, Vancouver. McCune, B., Grace, J.B., with a contribution from D.L. Urban, 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, Ore., USA. Meades, W.J., Moores, L., 1994. Forest Site Classification Manual: A Field Guide to the Damman Forest Types of Newfoundland, second ed. Western Newfoundland Model Forest, Inc., Dep. For. Agric., Nat. Resour. Can., Can. For. Serv., For. Resour. Dev. Agreement FRDA Rep. 003. Mohler, C.L., Marks, P.L., Sprugel, D.G., 1978. Stand structure and allometry of trees during self-thinning of pure stands. J. Ecol. 66, 599–614. Moloney, K.A., 1986. Wave and nonwave regeneration processes in a subalpine Abies balsamea forest. Can. J. Bot. 64, 341–349. Moreau, G., 2006. Past and present outbreaks of the balsam fir sawfly in western Newfoundland: an analytical review. For. Ecol. Manage. 221, 215–219. Morin, H., Laprise, D., 1990. Histoire re´cente des e´pide´mies de la Tordeuse des bourgeons de l’e´pinette au nord du lac Saint-Jean (Que´bec): une analyse dendrochronologique. Can. J. For. Res. 20, 1–8. Newton, P.F., 1990. Initial size hierarchy development within second-growth black spruce stands. Can. J. For. Res. 20, 1720–1731.

308

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309

Newton, P.F., 1992. Base-age invariant polymorphic site index curves for black spruce and balsam fir within central Newfoundland. N.J. Appl. For. 9, 18– 22. Newton, P.F., Smith, V.G., 1988. Diameter distributional trends within mixed black-spruce/balsam-fir and pure black-spruce stand types. For. Ecol. Manage. 25, 123–138. Newton, P.F., Jolliffe, P.A., 1998a. Assessing processes of intraspecific competition within spatially heterogeneous black spruce stands. Can. J. For. Res. 28, 259–275. Newton, P.F., Jolliffe, P.A., 1998b. Aboveground modular component responses to intraspecific competition within density-stressed black spruce stands. Can. J. For. Res. 28, 1587–1610. Ogawa, K., Hagihara, A., 2003. Self-thinning and size variation in a sugi (Cryptomeria japonica D. Don) plantation. For. Ecol. Manage. 174, 413– 421. Osawa, A., Allen, R.B., 1993. Allometric theory explains self-thinning relationships of mountain beech and red pine. Ecology 74, 1020–1032. Oshima, Y., Kimura, M., Iwaki, H., Kuroiwa, S., 1958. Ecological and physiological studies on the vegetation of Mt. Shimagare I. Preliminary survey of the vegetation of Mt. Shimagare. Bot. Mag. (Tokyo) 71, 289–301. Otvos, I.S., Moody, B.H., 1978. The spruce budworm in Newfoundland: History, status and control. Fish. Environ. Can., For. Serv., Newfoundland For. Res. Cen. Inf. Rep. N-X-150. Otvos, I.S., Clarke, L.J., Durling, D.S., 1979. A history of recorded eastern hemlock looper outbreaks in Newfoundland. Environ. Can., For. Serv., Newfoundland For. Res. Cent. Inf. Rep. N-X-179. Page, G., 1968. Site-index curves for spruce and fir in the forest regions of Newfoundland. Dep. For. Rural Dev., For. Br., For. Res. Lab. Inf. Rep. N-X22. Page, G., Bouzane, J.P., Bajzak, D., van Nostrand, R.S., 1970. A study of the mensurational characteristics of some important forest types of northern Newfoundland. Can. For. Serv, Dept. Fish. For., For. Res. Lab. Inf. Rep. NX-41. Peet, R.K., Christensen, N.L., 1980. Succession: a population process. Vegetatio 43, 131–140. Peet, R.K., Christensen, N.J., 1987. Competition and tree death. Bioscience 37, 586–595. Pickett, S.T.A., 1989. Space-for-time substitution as an alternative to long-term studies. In: Likens, G.E. (Ed.), Long-term Studies in Ecology: Approaches and Alternatives. Springer-Verlag, New York, pp. 210–235. Poore, M.E.D., 1962. The method of successive approximation in descriptive ecology. Adv. Ecol. Res. 1, 35–68. Porter, K.B., MacLean, D.A., Beaton, K.P., Upshall, J., 2001. New Brunswick permanent sample plot database (PSPDB v1.0): User’s guide and analysis. Can. For. Serv., Atlantic For. Cent. Inf. Rep. M-X-209E. Pothier, D., Savard, F., 1998. Actualisation des Tables de Production pour les Principales Espe`ces Forestie`res du Que´bec. Gouvernment du Que´bec, Ministe`re des Ressources naturelles, Foreˆt Que´bec, Que´bec. Quinn, G.P., Keough, M.J., 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, Cambridge. Raske, A.G., Clarke, L.J., Sutton, W.J., 1982. The status of the spruce budworm in Newfoundland in 1981. Environ. Can., Can. For. Serv., Newfoundland For. Res. Cen. Inf. Rep. N-X-210. Robertson, A., 1993. Wind and wave forests: a case study and implications for silviculture. In: Alden, J., Mastrantonio, J.L., Ødum, S. (Eds.), Forest Development in Cold Climates. Plenum Press. Published in cooperation with NATO Scientific Affairs Division, New York, pp. 491–510. Robichaud, E., Methven, I.R., 1993. The effect of site quality on the timing of stand breakup, tree longevity, and the maximum attainable height of black spruce. Can. J. For. Res. 23, 1514–1519. Rowe, J.S., 1977. Forest Regions of Canada. Can. For. Serv. Publ. No. 1300, reprint ed. Ottawa, Ontario. Sackville Hamilton, N.R., Matthew, C., Lemaire, G., 1995. In defence of the 3/2 boundary rule: a re-evaluation of self-thinning concepts and status. Ann. Bot. 76, 569–577. Smith, V.G., Watts, M., James, D.F., 1987. Mechanical stability of black spruce in the clay belt region of northern Ontario. Can. J. For. Res. 17, 1080–1091.

Solomon, D.S., Zhang, L., 2002. Maximum size–density relationships for mixed softwoods in the northeastern USA. For. Ecol. Manage. 155, 163– 170. Spies, T.A., Franklin, J.F., Thomas, T.B., 1988. Coarse woody debris in Douglas-fir forests of western Oregon and Washington. Ecology 69, 1689–1702. Sprugel, D.G., 1984. Density, biomass, productivity, and nutrient-cycling changes during stand development in wave-generated balsam fir forests. Ecol. Monogr. 54, 165–186. Sprugel, D.G., Bormann, F.H., 1981. Natural disturbance and the steady state in high-altitude balsam fir forests. Science 211, 390–393. SPSS Inc., 2002. Sigmaplot1 8.0 User’s Guide. Chicago, IL. Sturtevant, B.R., Bissonette, J.A., Long, J.N., 1998. Stand density management diagram for mixed balsam fir-black spruce stands. N.J. Appl. For. 15, 17–22. Sturtevant, B.R., Bissonette, J.A., Long, J.N., Roberts, D.W., 1997. Coarse woody debris as a function of age, stand structure, and disturbance in boreal Newfoundland. Ecol. Appl. 7, 702–712. Tadaki, Y., Sato, A., Sakurai, S., Takeuchi, I., Kawahara, T., 1977. Studies on the production structure of forest. XVIII. Structure and primary production in subalpine ‘‘dead tree strips’’ Abies forest near Mt. Asahi. Jpn. J. Ecol. 27, 83–90. Taylor, S.L., MacLean, D.A., 2005. Rate and causes of decline of mature and overmature balsam fir and spruce stands in New Brunswick, Canada. Can. J. For. Res. 35, 2479–2490. Thompson, I.D., Larson, D.J., Montevecchi, W.A., 2003. Characterization of old ‘‘wet boreal’’ forests, with an example from balsam fir forests of western Newfoundland. Environ. Rev. 11, S23–S46. Trexler, J.C., Travis, J., 1993. Nontraditional regression analysis. Ecology 74, 1629–1637. Warren, G.R., Meades, J.P., 1986. Wood defect and density studies. II. Total and net volume equations for Newfoundland forest management units. Can. For. Serv., Newfoundland For. Cen. Inf. Rep. N-X-242. Watkinson, A.R., Lonsdale, W.M., Firbank, L.G., 1983. A neighbourhood approach to self-thinning. Oecologia 56, 381–384. Weiner, J., 1984. Neighbourhood interference amongst Pinus rigida individuals. J. Ecol. 72, 183–195. Weiner, J., 1990. Asymmetric competition in plant populations. Trends Ecol. Evol. 5, 360–364. Weller, D.E., 1987a. A reevaluation of the 3/2 power rule of plant selfthinning. Ecol. Monogr. 57, 23–43. Weller, D.E., 1987b. Self-thinning exponent correlated with allometric measures of plant geometry. Ecology 68, 813–821. Weller, D.E., 1989. The interspecific size–density relationship among crowded plant stands and its implications for the 3/2 power rule of self-thinning. Am. Nat. 133, 20–41. West, P.W., Borough, C.J., 1983. Tree suppression and the self-thinning rule in a monoculture of Pinus radiata D. Don. Ann. Bot. 52, 149–158. West, P.W., Jackett, D.R., Borough, C.J., 1989. Competitive processes in a monoculture of Pinus radiata D. Don. Oecologia 81, 57–61. Westoby, M., 1984. The self-thinning rule. Adv. Ecol. Res. 14, 167–225. White, J., 1980. Demographic factors in populations of plants. In: Solbrig, O.T. (Ed.), Demography and Evolution in Plant Populations. Botanical Monographs, vol. 15. Blackwell Scientific Publications, Oxford, pp. 21–48. White, J., 1981. The allometric interpretation of the self-thinning rule. J. Theor. Biol. 89, 475–500. White, J., Harper, J.L., 1970. Correlated changes in plant size and number in plant populations. J. Ecol. 58, 467–485. Wilson, D.S., 1996. Theory and application of conifer growth dynamics based on plant allometry. M.S. thesis. University of Maine, p. 115. Wilson, D.S., Seymour, R.S., Maguire, D.A., 1999. Density management diagram for northeastern red spruce and balsam fir forests. N.J. Appl. For. 16, 48–56. Wyszomirski, T., 1992. Detecting and displaying size bimodality: kurtosis, skewness and bimodalizable distributions. J. Theor. Biol. 158, 109–128. Xue, L., Hagihara, A., 1999. Density effect, self-thinning and size distribution in Pinus densiflora Sieb. et Zucc. stands. Ecol. Res. 14, 49–58. Yarranton, M., Yarranton, G.A., 1975. Demography of a jack pine stand. Can. J. Bot. 53, 310–314.

J.W. McCarthy, G. Weetman / Forest Ecology and Management 241 (2007) 295–309 Yoda, K., Kira, T., Ogawa, H., Hozumi, K., 1963. Self-thinning in overcrowded pure stands under cultivated and natural conditions (Intraspecific competition among higher plants XI). J. Biol. Osaka City Univ. 14, 107–129. Zeide, B., 1987. Analysis of the 3/2 power law of self-thinning. For. Sci. 33, 517–537.

309

Zeide, B., 1991. Self-thinning and stand density. For. Sci. 37, 517– 523. Zhang, L., Bi, H., Gove, J.H., Heath, L.S., 2005. A comparison of alternative methods for estimating the self-thinning boundary line. Can. J. For. Res. 35, 1507–1514.