Cattle grazing facilitates tree regeneration in a conifer forest with palatable bamboo understory

Forest Ecology and Management 252 (2007) 73–83 www.elsevier.com/locate/foreco

Cattle grazing facilitates tree regeneration in a conifer forest with palatable bamboo understory A. Darabant a,b,*, P.B. Rai b, K. Tenzin a,c, W. Roder b,1, G. Gratzer a a

BOKU-University of Natural Resources and Applied Life Sciences, Department of Forest and Soil Sciences, Peter Jordan Str. 82, 1190 Vienna, Austria b Renewable Natural Resources Research Centre-Jakar, Bumthang, Bhutan c Renewable Natural Resources Research Centre-Yusipang, Thimphu, Bhutan Received 9 February 2007; received in revised form 6 June 2007; accepted 11 June 2007

Abstract In a temperate mixed conifer forest in the Bhutan Himalayas, we investigated the effects of cattle grazing on conifer (Tsuga dumosa, Pinus wallichiana, Picea spinulosa, Abies densa) seedling density, growth and microsite of recruitment under dense cover of the bamboo Yushania microphylla, using exclosures over a period of 9 years after group selection harvest. Increasing bamboo competition over time following canopy opening successfully prevented seedling recruitment in ungrazed plots, while recruitment in grazed plots was continuous. Reduction of bamboo height through grazing facilitated recruitment of all tree species, particularly T. dumosa, mainly through increased light interception on the forest floor. Tree species composition of seedlings and the overstory did not differ in ungrazed plots, while in grazed plots we observed a shift towards dominance of T. dumosa. Growth rates of T. dumosa and P. spinulosa were higher in grazed plots as compared to ungrazed plots. In grazed plots, recruitment of T. dumosa was concentrated on moss, which might have prevented desiccation of the small-seeded species after germination. We propose that controlled grazing might facilitate natural regeneration after logging in mixed conifer forests of central Bhutan with dense Y. microphylla bamboo understory. # 2007 Elsevier B.V. All rights reserved. Keywords: Cattle grazing; Bamboo; Forest regeneration; Facilitation; Arrested succession; Bhutan Himalayas

1. Introduction Large herbivores exert considerable influence on dynamics of ground vegetation (Milchunas and Lauenroth, 1993) and tree regeneration (Ross et al., 1970; Linhart and Whelan, 1980). Large herbivores can affect establishment and survival of tree seedlings either directly through browsing (Gill, 1992), trampling (Mitchell and Kirby, 1990), zoochorous seed dispersal and dispersal of mycorrhiza, or indirectly through soil compaction (Mitchell and Kirby, 1990), nutrient removal or input (Gill and Beardall, 2001), increased susceptibility to drought (Peterson and Pickett, 1995), change of competing

* Corresponding author at: BOKU-University of Natural Resources and Applied Life Sciences, Department of Forest and Soil Sciences, Peter Jordan Str. 82, 1190 Vienna, Austria. Tel.: +43 1 47654 4124; fax: +43 1 47654 4129. E-mail address: [email protected] (A. Darabant). 1 Present address: CIP/CFC, c/o National Potato Development Program, Simtokha, Thimphu, Bhutan. 0378-1127/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2007.06.018

understory vegetation (Kirby, 2001; Rossell et al., 2005), and possible resulting increases in rodent densities (Evans et al., 2006). With all other factors having mainly negative effects on tree seedling recruitment, changes in competing vegetation can be both, negative through increased dominance of nonpalatable competitors (de la Cretaz and Kelty, 1999; Royo and Carson, 2006), or positive through reduction of palatable competitors (Miles and Kinnaird, 1979). This indirect positive effect can outweigh direct negative effects on establishment and survival of tree seedlings and saplings (Gill and Beardall, 2001; Itoˆ and Hino, 2005). The effects of large herbivores on tree regeneration can differ with understory type (Peterson and Pickett, 1995; Nomiya et al., 2002; Kramer et al., 2006), herbivore density (Horsley et al., 2003; Mayer et al., 2006) and site conditions (Horsley et al., 2003). While no clear-cut recommendations exist for most forest types, moderate grazing has been proposed to facilitate tree regeneration under certain palatable understory vegetation types inhibiting tree regeneration (Kuiters et al., 1996; Miller and Wells, 2003; Mountford and Peterken, 2003; Pollock et al., 2005).

74

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

Dense understorys of bamboo species often create strong competition for light (Nakashizuka, 1988; Taylor and Qin, 1988; Cao, 1995), water (Takahashi et al., 2003), and nutrients (Numata, 1979), provide shelter for rodents resulting in increased seed predation and seedling damage (Wada, 1993; Abe et al., 2001; Iida, 2004) and thereby temporarily suppress tree regeneration in several locations world wide, e.g. Bhutan (Gratzer et al., 1999), China (Taylor and Qin, 1988; Wang et al., 2006), Japan (Nakashizuka and Numata, 1982; Nakashizuka, 1991), Chile (Veblen, 1982), Thailand (Marod et al., 1999), Costa Rica (Widmer, 1998), and Brazil (Guilherme et al., 2004). As a result of vegetative spread, fast growth response to increasing light, considerable tolerance against browsing, and depleting soil water resources, several bamboo species appear to form persistent understory layers, leading to delayed succession (Nakashizuka and Numata, 1982; Griscom and Ashton, 2003; Taylor et al., 2006; Royo and Carson, 2006). These understory layers act as a selective ecological filter, differentially affecting the regeneration of certain tree species (George and Bazzaz, 1999a; Wang et al., 2006). In the temperate zone of Bhutan, Yushania microphylla (Munro) R.B. Majumdar is a dominant and widespread, frequently browsed bamboo species, which is often associated with regeneration failure in harvested areas (Rosset, 1999). Virtually the entire forest area of Bhutan is grazed by domestic livestock, mostly through migratory herds of cattle and yaks (Gyamtsho, 2000; Norbu, 2002). The lack of forest regeneration in harvested areas in the country has mostly been attributed to the detrimental effects of cattle grazing, irrespective of understory type and grazing pressure (Miller, 1986; van Ijssel, 1990; Wangda and Ohsawa, 2006). Steps towards discouraging traditional grazing practices have met strong resistance from farmers, who heavily rely on livestock productivity (Ura, 2002). Roder et al. (2003) documented that without additional input of chemical fertilizers in agriculture, forest grazing is essential to maintain production levels through direct nutrient transfer from the forests to agricultural land, most notably of P. The reduction of bamboo height through moderate grazing was found to promote forest regeneration through increased light availability on the forest floor (Gratzer et al., 1999; Itoˆ and Hino, 2005). Therefore, moderate forest grazing by domestic herbivores might be beneficial in Bhutan, wherever justified from the silvicultural point of view (Gratzer et al., 1999). We conducted an experimental study comparing grazed and ungrazed plots over a period of 9 years to clarify the effects of cattle grazing on conifer regeneration under dense Y. microphylla bamboo understory in small size group openings in a temperate mixed conifer forest in central Bhutan. We had the specific objectives to (1) assess the impact of grazing on bamboo height and cover, (2) evaluate the tree species-specific effects of bamboo height reduction through grazing on tree seedling density, size distribution and growth, and (3) assess differences in site conditions affecting tree seedling establishment between grazed and ungrazed plots.

2. Methods 2.1. Study area The study was conducted at two sites in East-Central Bhutan in the Chumey valley of Bumthang district on rather gentle, east facing slopes at altitudes of 3225 m (Hurchi) and 3060 m (Domkhar) (Fig. 1). The area is dominated by meta-sediments and gneisses of the main crystalline belt of the high Himalaya (Gansser, 1983). Based on records of a nearby weather station (Hurchi, 3400 m), annual precipitation is 1300–1500 mm, the vast majority of which falls during monsoon from May to early October. The mean annual temperature is +4.6 8C, with the mean maximum temperature of +14.6 8C in July and the mean minimum temperature of 8.3 8C in January (Dorji, 2001; Bu¨rgi, 2002). The vegetation period (definition see Rinchen and Rosset (1996)) is approx. 200 days with early frost starting from the second week of October and late frost ending by May (Tenzin and Rinzin, 2003). The forest in the research area is dominated by Tsuga dumosa (D. Don) Eichler (Himalayan Hemlock) and Pinus wallichiana A.B. Jackson (Blue Pine), with the almost continuous presence of Picea spinulosa (Griff.) Henry (East Himalayan Spruce). Abies densa Griff. (East Himalayan Fir) occurs scattered in the research area, which is part of the transitional belt towards mono-specific fir forests occurring at higher altitudes. Stands, especially the ones dominated by pine are fairly young, without considerable amount of coarse woody debris. The understory is almost exclusively dominated by the bamboo Y. microphylla (Munro) R.B. Majumdar, sometimes mixed with Arundinaria racemosa Munro, another bamboo species. The biomass and height of both bamboo species are considerably reduced by cattle grazing. Shrubs, mostly Berberis aristata D.C., frequently occur in the area (Grierson and Long, 1983). Both, sedentary and migratory livestock graze in the study area. Migratory cattle, representing 50–70% of the livestock population in Chumey, are present for about 5 months every year during the summer. They migrate to lower lying areas in the south for winter and are replaced by a lower number of migratory yaks coming from pastures above the timberline. A livestock survey showed that Hurchi and Domkhar are grazed by comparable numbers of domestic ungulates (168 versus 194), with the numbers of sedentary livestock being virtually equal between the two locations (25 versus 23) (Dorji, 1997). Besides domestic ungulates, sambar (Cervus unicolor), wild boar (Sus scropha), and barking deer (Muntiacus muntjak) are common in the study area. Based on visual investigation of droppings, domestic livestock is present at higher densities than wild herbivores, on whose density no objective assessment has been carried out. 2.2. Experimental design and data recording In both locations, Hurchi and Domkhar, five pairs of plots were established in late 1996 in small group openings along cable line corridors created a few months earlier. Each pair of adjacent plots, representing a block, consisted of a randomly

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

75

Fig. 1. Map indicating location of Chumey in Bhutan (bottom right), the location of Hurchi and Domkhar research areas in Chumey (top right), locations of blocks in group openings along cable corridors in Hurchi (bottom left) and Domkhar incl. the design of a block with fenced and unfenced plots incl. subplots (top left).

assigned grazed and ungrazed plot of 4 m  6 m. Each plot was subdivided into six 2 m  2 m subplots to allow for fine-scaled assessment of regeneration, competing vegetation and microsites. Fencing was carried out after logging and initial recording in late 1996 (Tenzin and Rinzin, 2003). A buffer of 0.5 m was maintained on the perimeter of ungrazed plots in order to prevent bias through increased lateral light interception within the fence due to lower bamboo height outside the fence. The fence was designed to keep out large herbivores, both domestic and wild, allowing access for small herbivores. Our design did not allow to account for the possibly increased damage on tree seedling regeneration caused by small herbivores in the shelter of the fence. Even though this effect has been indicated by another study carried out in Hurchi (Rosset and Rinchen, 1998), our results did not confirm it (data not presented). Parameters recorded in early 1997 included number of conifer regeneration by species, bamboo height and cover percentage by layer, and microsite cover percentage. Parameters recorded in 2005 for conifer regeneration included species, height, root collar diameter, and microsite (moss, nurse log, bamboo litter, other litter, mineral soil, boulders and decomposed organic matter; recorded only for seedlings and saplings up to 50 cm height). At the same time, for each subplot we recorded microsite cover percentage, bamboo cover

percentage layer-wise and maximum bamboo height in four 1 m  1 m quadrants. The mean of the four maximum bamboo heights for each subplot was used for analysis. A hemispherical photograph was taken at 30 cm above the forest floor in the middle of each subplot to characterize light availability beneath the bamboo understory. Unless noted differently, all recordings refer to the subplot level. 2.3. Data analysis We used SPSS version 12.0.1 (SPSS Inc., 2003) for statistical analysis. We tested differences over time in seedling abundance and bamboo cover between grazed and ungrazed plots using repeated measures ANOVA (rmANOVA) (von Ende, 2001), having aggregated subplot data at the plot level. The between subject factors were block and treatment (ungrazed/grazed) and the within subject factors were time and the interactions of time with the between subject factors (Royo and Carson, 2005). When necessary, we normalized dependent variables through logarithmic transformation and tested residuals for normal distribution using the Shapiro– Wilks Test. Differences in tree seedling species distributions between grazed and ungrazed plots were tested using x2 goodness-of-fit tests. Expected values were based on overstory

76

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

tree composition of the boundary trees of the group openings. For this purpose, plots were divided into two groups, one group dominated by hemlock (plots 3–5), the other group dominated by pine with presence of hemlock (remaining plots). In the associated a posteriori analyses, we compared standardized residuals with critical values of the normal distribution to obtain significance levels (Lowry, 1998–2006; Jim Schwab, pers. comm.). We evaluated differences in size distributions between grazed and ungrazed plots using two-sample Kolmogorov– Smirnoff tests. We tested differences in relative growth rates with a three-factorial ANCOVAwith block, fencing and tree species as factors and fraction of solar irradiation as covariate. We used the relative growth rate of tree seedlings for 2005 (terminal shoot length 2005 [cm]/seedling height [cm]) as growth parameter to prevent bias due to the uneven distribution of size classes between grazed and ungrazed plots, tree species (Fig. 4B and C), as well as blocks. We tested differential survival of seedlings on certain microsites using x2 goodness-of-fit tests. A posteriori analyses were performed the way described above for x2 goodness-of-fit tests. We analyzed hemispherical photographs using Hemiview 2.1 (Delta-T Devices, 1998), estimating the proportion of solar radiation reaching the particular point where the photograph was taken = Global Site Factor (fraction of solar irradiation). We characterized the relationship between mean maximum bamboo height and fraction of solar irradiation using non-linear regression. The effect of bamboo cover on the number of seedlings was assessed using multiple stepwise regressions with the cover percentage of bamboo at ankle height (15 cm) (CBA), at knee height (45 cm) (CBK), at man height (175 cm), above man height (>175 cm), the maximum bamboo height [cm] (MBH) and fraction of solar irradiation (L) being the independent variables. 3. Results 3.1. Effect of fencing on bamboo cover and tree seedling density Mean tree seedling numbers in 1997 were 2792 ha1 in grazed plots and 4417 ha1 in ungrazed plots, the difference was however not significant (ANOVA, p > 0.05). In 2005, we recorded a mean seedling number of 19,125 ha1 in grazed

Fig. 2. Seedling density in 1997 and 2005 in grazed and ungrazed plots.

plots and 4583 ha1 in ungrazed plots (Fig. 2; Table 1). We found a significant increase of seedling density over time (rmANOVA, p  0.05). The significant treatment over time interaction ( p  0.01) showed that the increase of seedling numbers in grazed plots was greater than in ungrazed plots (Table 2; Fig. 2). Species-wise analysis of the data revealed that the increase in seedling numbers over time was largely attributed to an increased recruitment of hemlock in grazed plots (875 ha1 in 1997 versus 15,292 ha1 in 2005; Table 1). Stronger increase in grazed plots as compared to ungrazed plots was revealed through the significant fencing over time interactions for all tree species with the exception of fir, where no analysis could be performed due to low sample size (Table 2). Density of all tree species in ungrazed plots remained relatively constant over time (1997–2005: 1750 ha1 versus 1583 ha1 for hemlock, 333 ha1 versus 167 ha1 for spruce, 2250 ha1 versus 2583 ha1 for pine and 83 ha1 versus 250 ha1 for fir; Fig. 3; Table 1). Number of recruits larger than 25 cm in height did not differ between grazed and ungrazed plots in 2005 (Mann–Whitney U-Test, p > 0.05).

Table 1 Mean number of conifer seedlings per ha, bamboo cover (%), maximum bamboo height and fraction of solar irradiation in grazed and ungrazed plots in 1997 and 2005 including standard error of the mean Species

Grazed

Ungrazed

1997 (S.E.)

2005 (S.E.)

1997 (S.E.)

2005 (S.E.)

Conifer total

2792 (921)

19125 (9188)

4417 (1618)

4583 (2332)

Tsuga dumosa Picea spinulosa Pinus wallichiana Abies densa

875 (454) 125 (89) 1708 (728) 83 (83)

15292 792 2792 250

(8541) (487) (814) (208)

1750 (1258) 333 (247) 2250 (905) 83 (83)

1583 167 2583 250

Yushania microphylla cover (%) Y. microphylla max. height (cm) Fraction of solar irradiation (%)

22.3 (9.70) n.a. n.a.

64.0 (4.70) 121.0 (8.97) 0.300 (0.013)

19.6 (8.70) n.a. n.a.

(1020) (127) (1861) (208)

77.9 (2.90) 326.4 (10.12) 0.136 (0.009)

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

77

Table 2 Results of repeated-measures ANOVA for number of conifer regeneration and bamboo cover (A. densa has been omitted due to low sample size) Species

Between plot analysis over time

Within plot analyses

Fencing effects

Time effects

Fencing  time

d.f.

F

p

d.f.

d.f.

F

p

Conifer total

1,9

7.519

0.023

1,9

F 6.292

p 0.033

1,9

16.748

0.003

T. dumosa P. spinulosa P. wallichiana

1,9 1,9 1,9

5.944 1.876 3.735

0.037 0.204 0.085

1,9 1,9 1,9

7.923 2.024 0.359

0.020 0.189 0.564

1,9 1,9 1,9

7.770 6.119 11.873

0.021 0.035 0.007

Y. microphylla cover (%)

1,9

2.103

0.181

1,9

223.888

<0.0001

1,9

10.450

0.010

species composition corresponded to the overstory tree species composition in both hemlock- and pine-dominated openings. In grazed plots, hemlock was over-represented, while spruce and pine were under represented in both hemlock- and pinedominated openings (Table 3). The same tests conducted for seedlings larger than 20 cm in height yielded identical results. Plots located in hemlockdominated openings were the exception, where we found no significant deviation of tree seedling species composition from overstory tree species composition (results not presented). 3.3. Effect of fencing on seedling size distribution

Fig. 3. Seedling density by tree species in grazed and ungrazed plots in 1997 and 2005.

Mean bamboo cover also showed a highly significant increase over time, with a significant interaction of fencing over time (Table 2), confirming the stronger increase in ungrazed plots as compared to grazed plots (Table 1). Bamboo height was significantly higher in ungrazed plots compared to grazed plots in 2005 (Mann–Whitney U-Test, p  0.001). 3.2. Effect of fencing on tree seedling species composition We found a highly significant influence of grazing on tree seedling species composition. In ungrazed plots, tree seedling

We found significant differences between diameter and height distributions of seedlings between grazed and ungrazed plots (two-sample KS-Test, p  0.01). While differences in the upper ranges of the distributions appeared to be minor, substantial differences could be observed in smaller size classes, whereby small recruits were more abundant in grazed plots as compared to ungrazed plots (Fig. 4). A separate analysis for recruits larger than 25 cm in height showed no differences in diameter distribution (two-sample KS-Test, p > 0.05), but significant differences in height distribution (two-sample KS-Test, p  0.05) between grazed and ungrazed plots. Species-wise comparison of diameter distributions revealed that the great differences between grazed and ungrazed plots can mostly be attributed to hemlock. Smaller seedlings of hemlock were more abundant in grazed plots as compared to ungrazed plots, while this difference was not as strong with larger seedlings (Fig. 4B). In the case of pine this difference was relatively small and could only be observed in the smallest size classes (Fig. 4C). Differences in both distributions were highly significant (two-sample KS-Test, p  0.01). Separate analyses

Table 3 Observed vs. expected (in parentheses) seedling densities based on overstory tree species composition for plots with differing overstory composition, incl. significance of x2 tests of goodness-of-fit (A. densa omitted due to low sample size) Opening category

Gaps in P. wallichiana-dominated stands Gaps in T. dumosa-dominated stands

Grazed

Ungrazed 2

T. dumosa

P. spinulosa

P. wallichiana

x Sig.

T. dumosa

P. spinulosa

P. wallichiana

x2 Sig.

112 (23.6)** 255 (220.7)*

5 (50.1)** 14 (39.4)**

60 (103.3)** 7 (15.9)*

** **

10 (13.2) 28 (26.4)

4 (4.7)

61 (57.8) 1 (1.9)

n.s. n.s.

A posteriori tests are based on comparison of standardized residuals of the x2 tests with critical values of the normal distribution. Expected values are based on tree seedlings species distributions corresponding to species distribution of overstory trees. ** and * are significance levels for a posteriori tests.

78

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

for recruits larger than 25 cm in height showed significant differences in diameter and height distributions between grazed and ungrazed plots for hemlock (two-sample KS-Test, p  0.05), but no differences for pine (two-sample KS-Test, p > 0.05). 3.4. Effect of fencing on seedling growth Fencing and tree species both had a significant effect on relative growth rates of all seedlings combined (ANCOVA, p  0.05), while the interaction was not significant ( p = 0.120, Fig. 5). Block and the associated interactions were nonsignificant. Separate ANCOVAs for every tree species with block and fencing as factors and fraction of solar irradiation as covariate resulted in significant growth differences between grazed and ungrazed plots for hemlock and spruce, but not for pine and fir. In ungrazed plots, fir had the highest relative growth followed by pine and hemlock. Spruce had the lowest relative growth. In grazed plots, hemlock showed the highest relative growth followed by fir, spruce and pine (Fig. 5). 3.5. Species-wise microsite-specific survivorship response of tree seedlings Cover of different microsites in grazed plots differed from ungrazed plots (x2 test of goodness-of-fit, p  0.001). Fencing increased the proportion of bamboo litter microsite from 34% to 63%, proportionately decreasing the area of moss (24–17%) and other litter (31–15%). Other microsites had negligible cover (data not presented). Pine seedlings showed disproportionately low occurrence on bamboo litter and disproportionately high occurrence on moss,

Fig. 4. Diameter class distribution for grazed and ungrazed plots for all seedlings (A), Tsuga dumosa (B), and Pinus wallichiana (C).

Fig. 5. Relative growth rates of seedlings of four conifer species in grazed and ungrazed plots. Different letters indicate different group means ( p  0.05) within the same species.

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

79

Table 4 Observed vs. expected (in parentheses) values for seedlings on different microsites incl. significance of x2 tests of goodness-of-fit (nurse log omitted due to low seedling occurrence), and incl. significance of a posteriori tests based on standardized residuals of x2 tests compared with critical values of the normal distribution Species

T. dumosa P. spinulosa P. wallichiana A. densa

Grazed

Ungrazed

Moss

Other litter

Bamboo litter

x sig.

Moss

Other litter

Bamboo litter

x2 sig.

169 9 27 4

117 (116) n.s. 6 (7) 15 (19) n.s. 2 (2)

44 4 11 0

** n.s. **

7 0 28 0

1 0 2 0

13 3 10 6

n.s.

(89)** (5) (14)** (2)

(125)** (7) (20)* (2)

2

(4) (1) (7)** (1)

(3) (0) (6) n.s. (1)

(14) (2) (26)** (4)

**

No tests performed for P. spinulosa in ungrazed plots and A. densa due to low sample size. Expected values are based on even distribution of tree seedlings among microsites weighted by relative microsite cover. ** and * are significance levels for a posteriori tests.

regardless of fencing. Hemlock showed this characteristic only in grazed plots and occurred evenly distributed across all microsites in ungrazed plots. Spruce occurred evenly distributed across microsites in grazed plots, and no test could be performed in ungrazed plots due to low sample size. No test was performed for fir due to low sample size, however the trend indicated a slight preference for bamboo litter in ungrazed plots and moss in grazed plots (Table 4). 3.6. Effects of Y. microphylla height and cover on conifer regeneration Light levels on the forest floor dropped considerably with increasing bamboo height (Fig. 6). The relationship was modelled using a logarithmic function with r2 = 0.47. L ¼ 0:116  lnðmean maximum bamboo height ½cmÞ

Maximum bamboo height and seedling density were inversely related and maximum seedling numbers declined considerably above maximum bamboo heights of 0.75 m. Several subplots, regardless of maximum bamboo height showed no or low numbers of conifer regeneration (Fig. 7). Stepwise multiple regression models were developed showing the influence of bamboo cover percentage at ankle (CBA) and knee height (CBK), maximum bamboo height [cm] (MBH), and fraction of solar irradiation (L) on the total number of seedlings (NS) (r2 = 0.249), as well as on the number of hemlock (NH) (r2 = 0.238), and pine (NP) (r2 = 0.096), separately. NS ¼ 27:234  0:260ðCBKÞ  0:018ðHBÞ; NH ¼ 32:328  0:246ðCBKÞ  0:029ðHBÞ  22:648ðLÞ; NP ¼ 0:220 þ 6:051ðLÞ:

þ 0:798

Fig. 6. Scattergram and fitted regression line of mean maximum bamboo height and fraction of solar irradiation (Global Site Factor) 30 cm above the forest floor.

Fig. 7. Scattergram and boundary line of mean maximum bamboo height and tree seedling density per subplot.

80

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

4. Discussion 4.1. Effects of grazing on seedling density and tree species composition Indirect positive effects of grazing on tree seedling regeneration have been argued to outweigh direct negative effects under certain understory and grazing conditions (Evans et al., 2006; Gordon et al., 2004). While this does not apply in general (Kuiters and Slim, 2002), our study conforms to these findings. Itoˆ and Hino (2005), investigating differential effects of bamboo competition and grazing by large herbivores, reported positive indirect effects of browsing on tree regeneration through removal of the competing bamboo understory that has been found to inhibit tree seedling regeneration (Veblen et al., 1981; Nakashizuka and Numata, 1982; Nakashizuka, 1987; Taylor et al., 2004). Even though our study design did not allow for separate investigation of the factors grazing exclusion and bamboo competition, our findings show that their interactive effects negatively affect tree seedling establishment. While seedling numbers remained relatively constant in ungrazed plots, they increased more than fivefold in grazed plots over the course of nine vegetation periods (Fig. 2), indicating improved conditions for regeneration in grazed areas with bamboo cover. Herbivore density in our research area must have been low enough for indirect positive effects of grazing to outweigh direct negative ones (Itoˆ and Hino, 2005). The shift in tree seedling species composition over time, observed between grazed and ungrazed plots, highlights the selective filtering effect of the dense, tall bamboo understory on tree regeneration (Wang et al., 2006). Dense understory layers often act as an ecological filter influencing abundance, composition, size structure and spatial distribution of seedling banks mainly through competition for light (George and Bazzaz, 1999a,b; de la Cretaz and Kelty, 2002). Better survival of the small-seeded hemlock and spruce on grazed plots with reduced competition and larger cover of moss microsite protecting germinants from desiccation explain much higher seedling densities of these tree species in grazed plots. The shift in tree species composition, as observed with seedlings relative to the overstory tree composition in grazed plots (Table 3), indicates that grazing and/or bamboo dominance might not have existed at constant levels over the entire stand age and that they might exert an effect on future overstory tree composition (Ross et al., 1970; Tilghman, 1989; Wang et al., 2006), either by leading to an alternate stable state (Stromayer and Warren, 1997) or to altered gap-phase regeneration (Schnitzer et al., 2000). Increased recruitment of hemlock in pine-dominated forests supports the hypothesis that pine forests represent a successional stage towards hemlock-dominated mixed conifer forests in large areas of central Bhutan with formally heavy agricultural use (Rosset, 1999; Roder et al., 2002).

4.2. Effects of grazing on seedling size distribution and growth The unimodal size distributions of seedlings in ungrazed plots indicate increasingly reduced recruitment (Fig. 4), explained by increasing bamboo competition over time through lack of grazing (Tables 1 and 2). Reverse-J distributions of recruits in grazed plots show continuous establishment of seedlings and thus beneficial effects of competition control through grazing (Fig. 4). The lack of small-size recruits in ungrazed plots indicates that without grazing the dense bamboo understory might successfully prevent seedling establishment (Nakashizuka, 1988). More abundant tall recruits in ungrazed plots as compared to grazed plots possibly indicate increased browsing pressure near and above the general bamboo height (Hester et al., 2000; Pollock et al., 2005). Data on browsing damage only marginally supports this, as we found only 5.7% of recruits browsed in grazed plots (data not shown). The lower abundance of large size hemlock in grazed plots compared to ungrazed plots indicates greater preference of large ungulates for this tree species (Fig. 4B), which cannot be observed with pine (Fig. 4C). Seedling growth was considerably higher in grazed plots regardless of tree species (Fig. 5), mainly due to increased light interception with reduced bamboo height (Fig. 6) as a result of grazing (Table 2). This contradicts findings of other studies, which found reduced growth under grazed conditions for certain tree species (Ammer, 1996; Nomiya et al., 2002). In our study, the strong above- and below-ground competition, caused by the bamboo understory in ungrazed plots, and a low enough grazing pressure might have lead to higher growth rates under grazed conditions. The growth rate of hemlock corresponded to the findings of Gratzer et al. (2004), according to which the growth of hemlock between 14% and 30% of full sunlight mean fraction of solar irradiation in ungrazed and grazed plots (Table 1), should double. The lack of difference in growth rates of pine under shade and in the open confirms the low plasticity of the tree species, which cannot suppress its growth under shaded conditions and trades this off with higher mortality rates (Gratzer et al., 2004). Differences in growth rates of spruce between open and ungrazed conditions can be explained with the intermediate shade tolerance and growth plasticity of the tree species. The lack of difference in growth rates under shade and in the open with fir are surprising, given the high shade tolerance and plasticity of the tree species (Gratzer et al., 2004). Light levels in ungrazed plots might have been sufficiently high for fir to exhibit increased growth (Fig. 6), and fir has been shown to grow best on bamboo litter microsite in a study conducted 30 km away (Gratzer et al., 1999). 4.3. Effects of fencing on microsite distribution and microsite-specific survivorship response of seedlings Cattle and deer do not usually feed on bryophytes (Mitchell and Kirby, 1990), which can be one of the reasons for increased moss cover in grazed plots. Higher cover percentage of bamboo

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

was the reason for larger proportion of bamboo litter microsites in ungrazed plots. In temperate and subalpine conifer forests with Sasa sp. bamboo understory in Japan, recruitment of Picea jezoensis var. hondoensis and Tsuga diversifolia was substrate-associated with nurse logs, while Abies mariesii and A. veitchii regenerated regardless of substrate (Narukawa et al., 2003; Mori et al., 2004). Narukawa and Yamamoto (2002) showed that increasing cover of dwarf bamboo raises the relative importance of nurse logs for A. mariesii and A. veitchii regeneration. This was shown also for A. densa forests in Bhutan, where nurse logs were shown to be safe sites in the presence of Y. microphylla bamboo understory (Gratzer et al., 1999). As nurse logs, tip up mounds, etc. were hardly present in the relatively young stands we sampled, elevated microsites could not play a major role for regeneration in our study. Burial by litter (Christy and Mack, 1984) and competition with understory vegetation (Harmon and Franklin, 1989) might have played more important roles in determining the substrate of recruitment. Small patches of moss occurred in areas with lower bamboo density and thus lower cover of heavily decomposable bamboo litter (Narukawa and Yamamoto, 2002). Thick litter layers beneath a bamboo understory can act as a barrier for successful germination and establishment of particularly small-seeded tree species (Gonzales et al., 2002), while patches of moss provide shelter from desiccation (Maguire and Forman, 1983). Tree species with small seeds are expected to be more microsite-restricted than species with large seeds and thus greater resistance to various stresses (Grubb, 1977; Streng et al., 1989; Greene and Johnson, 1998). Small-seeded hemlock is prone to desiccation and therefore mostly occured on moss under open conditions, while it did not exhibit this characteristic under closed conditions, where the risk of desiccation is lower. The obligatory preference of pine for moss microsite regardless of fencing might be explained by the highly light demanding character of the tree species (Gratzer et al., 2004). Accordingly, pine seedlings show better survivorship response on moss patches because they occur in more open conditions. 4.4. Effect of bamboo competition on light availability and tree seedling density We showed that the improved resource availability for tree regeneration in grazed plots was mainly a result of lower competition for light by competing bamboo vegetation (Fig. 6), which has been reduced through grazing (Tables 1 and 2). Similarly to other understory types (Messier et al., 1989), stands of spreading bamboo have been found to darken the forest floor exponentially with their height, LAI or aboveground biomass (Gratzer et al., 1999; Itoˆ and Hino, 2004). The mean light levels we found under the closed bamboo canopy in ungrazed plots (13.6% of full sunlight) were well above 5%, 8%, 9%, and 12% of full sunlight, being the critical thresholds for increased mortality for hemlock, spruce, fir and pine, respectively (Gratzer et al., 2004). Dennstaedtia punctilobula, an understory fern species leading to arrested succession in the north-eastern part of the US, reduced light levels at the forest floor to 0.5% (de la Cretaz and Kelty, 2002) or 1.1% (George

81

and Bazzaz, 1999b) of full sunlight, successfully preventing tree regeneration. Mean maximum bamboo height appeared to influence maximum possible recruitment on the forest floor mainly, but not exclusively, through the light regime (Fig. 7). Other limiting factors might be competition for water (Takahashi et al., 2003), nutrients (Numata, 1979), increased seed and seedling predation (Abe et al., 2001), increased pathogen pressure and lack of preferred microsites under a dense bamboo cover (Table 4; Narukawa and Yamamoto, 2002). 4.5. Implications of bamboo grazing for forest dynamics and forest management Increased overstory disturbance together with increased herbivory have been proposed to lead to dominance of thick persistent understory layers (de la Cretaz and Kelty, 2002). These layers consist of a few plant species, which are usually light-demanding, but highly plastic with high growth rates, resistant to browsing/grazing and are often clonal, such as bamboo (Royo and Carson, 2006). Our findings in case of Y. microphylla in temperate conifer forests of central Bhutan show clear evidence for increased dominance of bamboo after overstory disturbance (Taylor et al., 1995), which in turn resulted in increased above-ground competition for light. Our results further indicate that increased litter accumulation under bamboo had a negative effect on conifer seedling establishment. Dense understory layers, such as bamboo, successfully delay tree seedling regeneration, primarily through competition for light (Lieffers et al., 1993; Abe et al., 2002; Taylor et al., 2006), and possibly alter successional pathways (Coomes et al., 2003) resulting in a change of canopy tree species composition (George and Bazzaz, 1999b; de la Cretaz and Kelty, 2002; Wang et al., 2006). We did not find evidence for regeneration failure, indicating that dominant layers of Y. microphylla allow for delayed forest regeneration, at least, when sufficiently controlled through grazing (Gratzer et al., 1999; de la Cretaz and Kelty, 1999). Controlled grazing with low enough livestock density might be a feasible option to promote regeneration after logging in conifer forests in Bhutan with dense, tall Y. microphylla bamboo understory. Acknowledgements The study was carried out by RNR-RC Jakar initially with support from Helvetas–Swiss Development Corporation, and later from BOKU University, Austria within the framework of the Conifer Research and Training Partnership funded by the Austrian Development Agency and the Royal Government of Bhutan (RGoB). We thank for the support by RGoB especially Dr. Pema Choephyel, N.K. Pradhan, Dr. Lungten Norbu, Dr. Tashi Dorji and Kinzang Wangdi. We would like to acknowledge the input by Rinchen and Jean Rosset at the start of the study. We are grateful to Tshewang Dorji, Sangay, Phurba Thinley, Karma Dorji and Pema Thinley for excellent support in the field and to Prof. Harald Strelec for statistical advice. We would also like to thank Tshering Dhendup with preparation of

82

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83

the map and the two anonymous reviewers who helped us in improving an earlier version of this manuscript. References Abe, M., Miguchi, H., Nakashizuka, T., 2001. An interactive effect of simultaneous death of dwarf bamboo, canopy gap and predatory rodents on beech regeneration. Oecologia 127, 281–286. Abe, M., Izaki, J., Miguchi, H., Masaki, T., Makita, A., Nakashizuka, T., 2002. The effects of Sasa and canopy gap formation on tree regeneration in an old beech forest. J. Veg. Sci. 13, 565–574. Ammer, C., 1996. Impact of ungulates on structure and dynamics of natural regeneration of mixed mountain forests in the Bavarian Alps. Forest Ecol. Manage. 88, 43–53. Bu¨rgi, A., 2002. Fir (Abies densa) forests in Central Bhutan: a model-based approach to assess a suitable utilization. Forestry 75, 457–464. Cao, K.-F., 1995. Fagus dominance in Chinese montane forests, natural regeneration of Fagus lucida and F. hayatae var. pashanica. Doctoral Thesis. Wageningen Agricultural University, The Netherlands. Christy, E.J., Mack, R.N., 1984. Variation in demography of juvenile T. heterophylla across the substratum mosaic. J. Ecol. 72, 75–91. Coomes, D.A., Allen, R.B., Forsyth, D.M., Lee, W.G., 2003. Factors preventing the recovery of New Zealand forests following control of invasive deer. Conserv. Biol. 17, 450–459. de la Cretaz, A.L., Kelty, M.J., 1999. Establishment and control of hay-scented fern: a native invasive species. Biol. Invasions 1, 223–236. de la Cretaz, A.L., Kelty, M.J., 2002. Development of tree regeneration in ferndominated forest understories after reduction of deer browsing. Restor. Ecol. 10, 416–426. Delta-T Devices Ltd., 1998. Hemiview User Manual 2.1. Delta-T Devices Ltd., Cambridge, UK. Dorji, K., 1997. Interview report on cattle grazing in two FMUs (Chumey valley). RNR-RC Jakar unpubl. Dorji, K., 2001. The weather in Bumthang. Retrieval http://www.moa.gov.bt/ rcjakar/f02.htm (accessed December 6, 2006; 15:40 GMT). Evans, D.M., Redpath, S.M., Elston, D.A., Evans, S.A., Mitchell, R.J., Dennis, P., 2006. To graze or not to graze? Sheep, voles, forestry and nature conservation in the British uplands. J. Appl. Ecol. 43, 499–505. Gansser, A., 1983. Geology of the Bhutan Himalaya (Denkschrifen der Schweizerischen Naturforschenden Gesellschaft), vol. 96. Birkha¨user, Basel, Switzerland, 181 pp. George, L.O., Bazzaz, F., 1999a. The fern understory as an ecological filter: emergence and establishment of canopy-tree seedlings. Ecology 80, 833– 845. George, L.O., Bazzaz, F., 1999b. The fern understory as an ecological filter: growth and survival of canopy tree seedlings. Ecology 80, 846–856. Gill, R.M.A., 1992. A review of damage by mammals in north temperate forests. 3. Impact on trees and forests. Forestry 65, 363–388. Gill, R.M.A., Beardall, V., 2001. The impact of deer on woodlands: the effects of browsing and seed dispersal on vegetation structure and composition. Forestry 74, 209–218. Gonzales, M.E., Veblen, T.T., Donoso, C., Valeria, L., 2002. Tree regeneration responses in lowland Nothofagus-dominated forests after bamboo dieback in South-Central Chile. Plant Ecol. 161, 59–73. Gordon, I.J., Hester, A.J., Festa-Bianchet, M., 2004. The management of wild large herbivores to meet economic, conservation and environmental objectives. J. Appl. Ecol. 41, 1021–1031. Gratzer, G., Rai, P.B., Glatzel, G., 1999. The influence of the bamboo Yushania microphylla on regeneration of Abies densa in central Bhutan. Can. J. Forest Res. 29, 1518–1527. Gratzer, G., Darabant, A., Rai, P.B., Chhetri, P., Eckmu¨llner, O., 2004. Interspecific variation in the response of growth, crown morphology, and survivorship to light of six tree species in the conifer belt of the Bhutan Himalayas. Can. J. Forest Res. 34, 1093–1107. Greene, D.F., Johnson, E.F., 1998. Seed mass and early survivorship of tree species in upland clearings and shelterwoods. Can. J. Forest Res. 28, 1307– 1316.

Grierson, A.J.C., Long, D.G., 1983. Flora of Bhutan, vol. 1, Part 1. Royal Botanic Garden, Edinburgh. Griscom, B.W., Ashton, P.M.S., 2003. Bamboo control of forest succession: Guadua sarcocarpa in Southeastern Peru. Forest Ecol. Manage. 175, 445– 454. Grubb, P.J., 1977. The maintenance of species-richness in plant communities: the importance of regeneration niche. Biol. Rev. Camb. Phil. Soc. 52, 107– 145. Guilherme, F.A.G., Oliviera-Filho, A.T., Appolina´rio, V., Bearzoti, E., 2004. Effects of flooding regime and woody bamboos on tree community dynamics in a section of tropical semideciduous forest in south-eastern Brazil. Plant Ecol. 174, 19–36. Gyamtsho, P., 2000. Economy of yak herders. J. Bhutan Stud. 2 (1) 45 p. Harmon, M.E., Franklin, J.F., 1989. Tree seedlings on logs in Picea-Tsuga forests of Oregon and Washington. Ecology 70, 48–59. Hester, A.J., Edenius, L., Buttenschon, R.M., Kuiters, A.T., 2000. Interactions between forests and herbivores: the role of controlled grazing experiments. Forestry 73, 381–391. Horsley, S.B., Stout, S.L., DeCalesta, D.S., 2003. White-tailed deer impact on the vegetation dynamics of a northern hardwood forest. Ecol. Appl. 13, 98– 118. Iida, S., 2004. Indirect negative influence of dwarf bamboo on survival of Quercus acorn by hoarding behaviour of wood mice. Forest Ecol. Manage. 202, 257–263. Itoˆ, H., Hino, T., 2004. Effects of mice and dwarf bamboo on the emergence, survival and growth of Abies homolepis (Pinaceae) seedlings. Ecol. Res. 19, 217–223. Itoˆ, H., Hino, T., 2005. How do deer affect tree seedlings on a dwarf bamboodominated forest floor? Ecol. Res. 20, 121–128. Kirby, K.J., 2001. The impact of deer on the ground flora of British broadleaved woodland. Forestry 74, 219–229. Kramer, K., Groot Bruinderink, G.W.T.A., Prins, H.H.T., 2006. Spatial interactions between ungulate herbivory and forest management. Forest Ecol. Manage. 226, 238–247. Kuiters, A.T., Slim, P.A., 2002. Regeneration of mixed deciduous forest in a Dutch forest-heathland, following a reduction of ungulate densities. Biol. Conserv. 105, 65–74. Kuiters, A.T., Mohren, G.M.J., van Wieren, S.E., 1996. Ungulates in temperate forest ecosystems. Forest Ecol. Manage. 88, 1–5. Lieffers, V.J., Macdonald, S.E., Hogg, E.H., 1993. Ecology of and control strategies for Calamagrostis canadensis in boreal forest sites. Can. J. Forest Res. 23, 2070–2077. Linhart, Y.B., Whelan, R.J., 1980. Woodland regeneration in relation to grazing and fencing in Coed Gorswen, North Wales. J. Appl. Ecol. 17, 827–840. Lowry, R., 1998–2006. Chi-square goodness of fit test. Retrieval http://faculty.vassar.edu/lowry/csfit.html (accessed December 6, 2006; 15:30 GMT). Maguire, D.A., Forman, R.T.T., 1983. Herb-cover effect on tree seedling patterns in a mature hemlock-hardwood forest. Ecology 64, 1367–1380. Marod, D., Kutintara, U., Yarwudhi, C., Tanaka, H., Nakashizuka, T., 1999. Structural dynamics of a natural mixed deciduous forest in western Thailand. J. Veg. Sci. 10, 777–786. Mayer, A.C., Sto¨ckli, V., Konold, W., Kreuzer, M., 2006. Influence of cattle stocking on browsing of Norway spruce in subalpine wood pastures. Agroforest. Syst. 66, 143–149. Messier, C., Honer, T.W., Kimmins, J.P., 1989. Photosynthetic photon flux density, red to far-red ratio, and minimum light requirement for survival of Gaultheria shallon in western red cedar-western hemlock stands in coastal British Columbia (Canada). Can. J. Forest Res. 19, 1470–1477. Milchunas, D.G., Lauenroth, W.K., 1993. Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecol. Monogr. 64, 327–366. Miles, J., Kinnaird, J.W., 1979. The establishment and regeneration of birch, juniper and Scots pine in the Scottish Highlands. Scott. Forest. 33, 102–119. Miller, C., Wells, A., 2003. Cattle grazing and the regeneration of totara (Podocarpus totara var. waihoensis) on river terraces, south Westland, New Zealand. NZ J. Ecol. 27, 37–44. Miller, D.J., 1986. Consultancy in Grazing Control, Forest Industries Complex, (BHU/80/014). Department of Forests, Thimphu.

A. Darabant et al. / Forest Ecology and Management 252 (2007) 73–83 Mitchell, F.J.G., Kirby, K.J., 1990. The impact of large herbivores on the conservation of semi-natural woods in the British Uplands. Forestry 63, 333–353. Mori, A., Mizumachi, E., Osono, T., Doi, Y., 2004. Substrate-associated seedling recruitment and establishment of major conifer species in an old-growth subalpine forest in central Japan. Forest Ecol. Manage. 196, 287–297. Mountford, E.P., Peterken, G.F., 2003. Long-term change and implications for the management of wood-pastures: experience over 40 years from Denny Wood, new forest. Forestry 76, 19–43. Nakashizuka, T., 1987. Regeneration dynamics of beech forests in Japan. Vegetatio 69, 169–175. Nakashizuka, T., 1988. Regeneration of beech (Fagus crenata) after the simultaneous death of undergrowing dwarf bamboo (Sasa kurilensis). Ecol. Res. 3, 21–35. Nakashizuka, T., 1991. Population dynamics of coniferous and broad-leaved trees in a Japanese temperature mixed forest. J. Veg. Sci. 2, 413–418. Nakashizuka, T., Numata, M., 1982. Regeneration process of climax beech forests. I. Structure of a beech forest with the undergrowth of Sasa. Jpn. J. Ecol. 32, 57–67. Narukawa, Y., Yamamoto, S., 2002. Effects of dwarf bamboo (Sasa sp.) and forest floor microsites on conifer seedling recruitment in a subalpine forest, Japan. Forest Ecol. Manage. 163, 61–70. Narukawa, Y., Iida, S., Tanouchi, A., Abe, S., Yamamoto, S.-I., 2003. State of fallen logs and the occurrence of conifer seedlings and saplings in boreal and subalpine old-growth forests in Japan. Ecol. Res. 18, 267–277. Nomiya, H., Suzuki, W., Kanazashi, T., Shibata, M., Tanaka, H., Nakashizuka, T., 2002. The response of forest floor vegetation and tree regeneration to deer exclusion and disturbance in a riparian deciduous forest, central Japan. Plant Ecol. 164, 263–276. Norbu, L., 2002. Grazing management in broadleaf forests. J. Bhutan Stud. 7 31 p. Numata, M., 1979. The relationship of limiting factors to the distribution and growth of Bamboo. In: Numata, M. (Ed.), Ecology of Grasslands and Bamboolands in the World. VEB Gustav Fisher, Jena, pp. 258–275. Peterson, C.J., Pickett, S.T.A., 1995. Forest reorganization—a case study in an old-growth forest catastrophic blowdown. Ecology 76, 763–774. Pollock, M.L., Milner, J.M., Waterhouse, A., Holland, J.P., Legg, C.J., 2005. Impacts of livestock in regenerating upland birch woodlands in Scotland. Biol. Conserv. 123, 443–452. Rinchen, Rosset, J., 1996. The weather in Bumthang 1996. RNR-RC Jakar, Bumthang. Retrieval http://www.moa.gov.bt/rcjakar/f02.htm (accessed May 30, 2007; 12:31 GMT). Roder, W., Gratzer, G., Wangdi, K., 2002. Cattle grazing in the conifer forests of Bhutan. Mountain Res. Develop. 22, 1–7. Roder, W., Dorji, K., Gratzer, G., 2003. Nutrient flow from the forest—source of life for traditional Bhutanese agriculture. Austrian J. Forest Sci. 120, 65–72. Ross, B.A., Bray, J.R., Marshall, W.H., 1970. Effects of long-term deer exclusion on a Pinus resinosa forest in north-central Minnesota. Ecology 51 (6), 1088–1093. Rossell, C.R., Gorsira, B., Patch, S., 2005. Effects of white-tailed deer on vegetation structure and woody seedling composition in three forest types on the Piedmont Plateau. Forest Ecol. Manage. 210, 415–424. Rosset, J., 1999. The temperate conifer forests of Bhutan. A review of forestry research activities until June 1998. Renewable Natural Resource Research Centre (RNR-RC) Jakar, Bumthang, Bhutan. Spec. Pub. 3, 95 pp. Retrieval http://www.moa.gov.bt/rcjakar/f02.htm (accessed May 30, 2007; 12:33 GMT). Rosset, J., Rinchen, 1998. Study of natural regeneration in small openings created by group selection silvicultural system in mixed conifer stands rich in Hemlock, central Bhutan. Renewable Natural Resource Research Centre (RNR-RC) Jakar, Bumthang, Bhutan. Retrieval http://www.moa.gov.bt/ rcjakar/f02.htm (accessed May 29, 2007; 16:01 GMT).

83

Royo, A.A., Carson, W.P., 2005. The herb community of a tropical forest in central Panama´: dynamics and impact of mammalian herbivores. Oecologia 145, 66–75. Royo, A.A., Carson, W.P., 2006. On the formation of dense understory layers in forests worldwide: consequences and implications for forest dynamics, biodiversity, and succession. Can. J. Forest Res. 36, 1345–1362. Schnitzer, S.A., Dalling, J.W., Carson, W.P., 2000. The impact of lianas of tree regeneration in tropical forest canopy gaps: evidence for an alternative pathway of gap-phase regeneration. J. Ecol. 88, 655–666. SPSS Inc., 2003. SPSS for Windows Release 12.0.1. SPSS Inc., Chicago, IL. Streng, D.R., Glitzenstein, J.S., Harcombe, P.H., 1989. Woody seedling dynamics of an East Texas floodplain forest. Ecol. Monogr. 59, 177–204. Stromayer, K.A.K., Warren, R.J., 1997. Are overabundant deer herds in the eastern United States creating alternate stable states in forest plant communities? Wildlife Soc. Bull. 25, 227–234. Takahashi, K., Uemura, S., Suzuki, J.-I., Hara, T., 2003. Effects of understory dwarf bamboo on soil water and the growth of overstory trees in a dense secondary Betula ermannii forest, northern Japan. Ecol. Res. 18, 767–774. Taylor, A.H., Qin, Z., 1988. Regeneration patterns in old-growth Abies-Betula forests in the Wolong Natural Reserve, Sichuan, China. J. Ecol. 76, 1204– 1218. Taylor, A.H., Qin, Z., Liu, J., 1995. Tree regeneration in an Abies faxoniana forest after bamboo dieback, Wang Lang Natural Reserve, China. Can. J. Forest Res. 25, 2034–2039. Taylor, A.H., Jinyan, H., ShiQiang, Z., 2004. Canopy tree development and undergrowth bamboo dynamics in old-growth Abies-Betula forests in southwestern China: a 12-year study. Forest Ecol. Manage. 200, 347–360. Taylor, A.H., Jang, S.W., Zhao, L.J., Liang, C.P., Miao, C.J., Huang, J., 2006. Regeneration patterns and tree species coexistence in old-growth AbiesPicea forests in southwestern China. Forest Ecol. Manage. 223, 303–317. Tenzin, K., Rinzin, 2003. Impact of livestock grazing on the regeneration of some major species of plants in conifer forest. Retrieval http://www.moa.gov.bt/rcjakar/f02.htm (accessed December 15, 2006; 12:40 GMT). Tilghman, N.G., 1989. Impacts of white-tailed deer on forest regeneration in northwestern Pennsylvania. J. Wildlife Manage. 53, 524–532. Ura, K., 2002. The herdsmen’s dilemma. J. Bhutan Stud. 7 (43). van Ijssel, W., 1990. The effect of grazing and seedbed type on forest regeneration under two silvicultural systems in mixed broadleaf forest at Gedu. Tsenden 3, 32–44. Veblen, T.T., 1982. Growth pattern of Chusquea bamboos in the understory of Chilean Nothofagus forests and their influences in the forest dynamics. Bull. Torrey Bot. Club 109, 474–487. Veblen, T.T., Donoso, C., Schlegel, F.M., Escobar, B., 1981. Forest dynamics in south-central Chile. J. Biogeogr. 8, 211–247. von Ende, C.N., 2001. Repeated-measures analysis: growth and other timedependent measures. In: Schreiner, S.M., Gurevitch, J. (Eds.), Design and Analysis of Ecological Experiments. Oxford University Press, NY, pp. 134– 157. Wada, N., 1993. Dwarf bamboos affect the regeneration of zoochorous trees by providing habitats to acorn-feeding rodents. Oecologia 94, 403–407. Wang, W., Franklin, S.B., Ren, Y., Oulette, J.R., 2006. Growth of bamboo Fargesia qinlingensis and regeneration of trees in a mixed hardwoodconifer forest in the Qinling Mountains, China. Forest Ecol. Manage. 234, 107–115. Wangda, P., Ohsawa, M., 2006. Structure and regeneration dynamics of dominant tree species along altitudinal gradient in a dry valley slopes of the Bhutan Himalaya. Forest Ecol. Manage. 230, 136–150. Widmer, Y., 1998. Pattern and performance of understory bamboos (Chusquea spp) under different canopy closures in old-growth oak forests in Costa Rica. Biotropica 30, 400–415.