Developmental dynamics following selective logging of an evergreen oak forest in the Eastern Himalaya, Bhutan: Structure, composition, and spatial pattern

Forest Ecology and Management 336 (2015) 163–173

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

Developmental dynamics following selective logging of an evergreen oak forest in the Eastern Himalaya, Bhutan: Structure, composition, and spatial pattern Kristofer Covey a,⇑, Charles J.W. Carroll b, Marlyse C. Duguid a, Kuenzang Dorji c, Tsewang Dorji d, Sonam Tashi e, Thinley Wangdi e, Mark Ashton a a

School of Forestry and Environmental Studies, Yale University New Haven, CT 06511, USA Graduate Degree Program in Ecology, Colorado State University, Fort Collins, CO 80523, USA c Ugyen Wangchuck Institute for Conservation and Environment, Lamai Gompa, Bumthang, Bhutan d Renewable Natural Resources Research and Development Center, Jakar, Bhutan e Renewable Natural Resources Research and Development Center, Yusipang, Bhutan b

a r t i c l e

i n f o

Article history: Received 8 April 2014 Received in revised form 25 September 2014 Accepted 1 October 2014

Keywords: Quercus semecarpifolia. karsu oak Selective logging Spatial pattern Old-growth Ripley’s K

a b s t r a c t Brown oak (Quercus semecarpifolia, a.k.a. Kharsu, bji shing) is a biologically and economically important evergreen broadleaved tree that dominates moist temperate and lower-montane forests throughout the mid-elevation Himalaya. We demarcated two paired spatially explicit one-hectare plots in an experimentally harvested area and an unharvested old growth reserve of Q. semecarpifolia dominated forest in the Bhutan Himalaya. We compared the structure, species composition and diversity, and spatial relationships between the two plots. To test whether harvesting had been successful in establishing a new cohort of oak we compared regeneration in plots established in 1999, to data gathered over ten years after. Regeneration plots showed a paucity of Quercus regeneration in both stands. Logging did not reduce tree species richness; however, Shannon diversity, Simpson diversity, and evenness were all lower in the logged stand. We used univariate and bivariate Ripley’s-K functions to assess the spatial distribution of trees in both stands and test whether single tree felling had altered the spatial relationships among and between species. Understory species were clumped at scales >30 m in canopy gaps in the old-growth reserve, whereas distribution in the logged plot was more random. Relationships between species show similar patterns with more than 80% of species showing significant clumping at scales from 12 m to 30 m, while 70% of relationships in the logged plot showed complete spatial randomness. In the old-growth reserve several species showed significant dispersion away from canopy dominant oaks. Scarce regeneration and significant changes in spatial pattern development in the harvested stand suggest changes to current silvicultural practice are needed. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction Existing at elevations between 2000 and 3000 m, Quercus semecarpifolia (brown, oak a.k.a. Kharsu, bji shing) is found throughout the moist mid-elevation Himalaya in pure stands and in mixed communities with broadleaved (e.g. Betula spp., Pyrus spp., Juglians regia, Prunus spp., Acer spp., Fraxinus spp., Rhododendron spp.), and conifer (e.g. Tsuga dumosa, Taxus baccata, Pinus wallichiana) tree species, and often with a diverse shrub layer (e.g. Rosa spp., Rubus spp., Virbunum spp., Lonicera spp. Pieris spp., Berberis spp. and Daphne bholua (Troup, 1921; Sargent, 1985; RGOB, 2004). Capable ⇑ Corresponding author. E-mail address: [email protected] (K. Covey). http://dx.doi.org/10.1016/j.foreco.2014.10.006 0378-1127/Ó 2014 Elsevier B.V. All rights reserved.

of withstanding minimum temperatures of 15 °C, Q. semecarpifolia prefers mean annual temperatures between 5 and 17 °C, annual rainfall around 1000–2500 mm, and a dry season not extending more than 4–6 months (Orwa et al., 2009). The species is most dominant on north-facing slopes from 2400 to 3600 m though in China it grows right up to the tree line, where it becomes a thicket-forming shrub (Gamble, 1881; Bean, 1916). At maturity Q. semecarpifolia commonly reaches 24–30 m tall and 60–70 cm in diameter at breast height (dbh). Seedling growth rates are modest, averaging 5–10 cm a year (Troup, 1921). In youth, the species is tolerant to moderate side shade (Huxley et al., 1992), but mature individuals are less tolerant (Gamble, 1881). Recognizing the impacts of deforestation in neighboring countries, the Bhutanese government has set as its goal the permanent

164

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

maintenance of 60% forest cover. Forests are to be managed for multiple values including: timber and non-timber forest products, biodiversity, water quality, grazing, fodder production, and human happiness (RGOB, 2004). Similar to other forest types in Bhutan, Q. semecarpifolia dominated forests have not been systematically managed in the past, and in addition are heavily grazed by cattle in the summer months, and yaks in the winter (Biswas, 1986; Davidson, 2000). Historically, silviculture has been limited to single tree removal, where canopy dominant Q. semecarpifolia are manually felled and removed to open gaps to promote understory development. Attempts at establishing natural regeneration of oak species in Bhutanese broadleaved forests have thus far proved unsuccessful, and there is growing concern about the suitability of this model for maintaining oak species following harvest (RGOB, 2004; Buffum et al., 2008) under current harvest regimes. Regeneration difficulties are not limited to Bhutan. In fact, most studies report a paucity of Q. semecarpifolia regeneration (Upreti et al., 1985; Singh and Singh, 1986; Dhar et al., 1997; Vetaas, 2000; Shrestha, 2003). Many of these studies suggest that intensive grazing is the most important factor inhibiting regeneration in Himalayan oak forests. However, oak’s recalcitrant, short-lived seeds and shade intolerance are often cited as limiting vigorous regeneration as well (Bisht et al., 2012; Verma et al., 2012). It is suspected that abundant light is the essential element in securing dense regeneration with areas near forest edges and isolated parent trees well stocked with vigorous seedlings (Jackson, 1984; Orwa et al., 2009). Metz (1997) also proposed that regeneration for this forest type is limited by small gap size and a lack of catastrophic disturbances (e.g. fire or windthrow), and suggests canopy openings greater than 400 m2 are necessary to regenerate existing stands, though ample regeneration has been observed in relatively undisturbed stands (Wangda and Ohsawa, 2006). Irregular seed production, defoliation, acorn predation, decreased or increased fire incidence and extensive lopping have also been blamed for scarce regeneration (Singh and Singh, 1986; Lorimer et al., 1994; Thadani and Ashton, 1995). Evidence from many oak forest dynamics studies from the past three decades suggest that allogenic disturbance and initial floristics restrict the maturation and development of most oak forest systems worldwide (Oliver, 1992; Abrams et al., 1995; Baker et al., 2005). These processes are well-demonstrated in other temperate deciduous and evergreen oak systems in Asia (Sano, 1997; Suh and Lee, 1998; Wangda and Ohsawa, 2006), Europe (Reif et al., 1999; Harmer and Morgan, 2007; Dobrowolska, 2008), and North America (Oliver and Stephens, 1977; Ruffner and Abrams, 1998; Liptzin and Ashton, 1999). Taken together, these studies describe a genus that is generally long-lived and intolerant of shade. Even where it dominates the canopy Quercus species are generally not present in lower strata, except in the aftermath of semi-lethal disturbance (Johnson et al., 2002). This literature suggests, a disconnect between the natural regeneration patterns of Quercus dominated forests and the single tree selection system practiced for rural use harvesting in Bhutanese forest management units. Although Q. semecarpifolia forests are important both culturally and economically, relatively little is known about them. There have been a few studies classifying the forest type (Sargent, 1985; Upreti et al., 1985), and exploring the social values of these forests (Fischer, 1976; Biswas, 1986), but to date there has been no work exploring spatial pattern, structure and composition of old-growth Q. semecarpifolia forests in relation to selective timber harvesting. In this study we examine the hypothesis that single-tree harvesting may not provide conditions adequate for the regeneration of Q. semecarpifolia. In addition, we test if selection harvesting alters the spatial relationships between and among species. We describe the composition and structure of harvested and unharvested

large-spatially explicit forest plots in a Bhutanese evergreen oak forest. We analyze differences in spatial patterns, and discuss implications for the sustainable management of Q. semecarpifolia dominated forests in the mid-elevation Himalaya. We believe, given the dearth of information within the Central Himalayan region that this is the first careful analysis of selective timber harvesting and its effects on this forest type’s spatial pattern, structure, regeneration, and composition.

2. Materials and methods 2.1. Site description In 1999 the Bhutanese Department of Forestry developed a 9 ha Q. semecarpifolia research area in Chimithanka, within the Gidikom Forest Management Unit (FMU) (27°260 N, 89°300 E), 15 km west of Thimphu (Fig. 1). Set at an elevation of 3000 m on a northwesterly aspect, slopes range from 20–70° with an average slope of 25°. Annual rainfall is approximately 720 mm per year and strongly seasonal, with the majority of rainfall concentrated during the summer monsoon months from June through August. Annual temperature varies between 8 °C and 30 °C. Soils throughout the site are fairly uniform; deep, acidic, well drained, and rich in organic matter (>5.5%) (BSS/NSSC, 2003); using the USDA (1975) soil classification the soils would be considered udults. The forest is routinely grazed in both the winter and summer months; evidence of browse is nearly ubiquitous. Stems of Q. semecarpifolia in Gidakom FMU are generally marked for removal by a government forester on behalf of a rural villager, felled by local workmen and removed by being rolled down hill, most commonly for use as fuelwood. Experimental felling and the establishment of understory regeneration plots took place in 2000 as part of an earlier study (Tashi and Thinley, 2008). At that time, a total of 12 large trees (9 Q. semecarpifolia. 1 Quercus thompsonii, 1 Betula alnoides, 1 T. dumosa) (11 m2 ha1, or 22% of total basal area) were removed from within the 1 ha harvested plot. Small diameter fuelwood was also removed from the stand at this time. Selection thinnings (as defined by Smith et al., 1997), like this one, where the largest stems of marketable species are removed are commonly prescribed for rural use harvesting in this forest type (RGOB, 2004). A network of 1142 m2 systematic regeneration plots were established on a 30 m grid in the harvested stand at this time. Additionally 1210  10 m plots were established following harvest to assess the long-term impacts of cattle and yak grazing in the stand. Located near the center of the stand, across a range of existing light environments. Half of these plots were fenced with 5-strand barbed wire, while the remaining 6 were left open to grazing. A detailed report of the existing regeneration and grazing exclusion plot network and a report of individual seedling performance over the 3 years following harvest was produced by Tashi and Thinley (2008). An adjacent 10 ha area of similar density (45 m2 ha1 of total basal area in the reserve vs. 49 m2 ha1 in the harvested) and composition (32.8 m2 ha1 Q. semecarpifolia in the reserve stand vs. 29.8 m2 ha1 in the harvested stand) was selected prior to harvest and set aside as an old-growth reserve to serve as a control stand prior to harvesting In June 2009, we demarcated and measured 1-hectare spatially explicit plots within paired, experimentally-harvested, and reserve stands. Due to extreme slopes (20–65%), spatially explicit plots were shaped differently to cover similar elevation gradients. The spatially explicit plot within the harvested stand was 50 m  200 m, and the spatially explicit plot within the reserve stand was 100 m  100 m. Both spatially explicit plots were randomly located inside an existing systematic network of

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

165

Fig. 1. Study Site.

regeneration plots, and are representative of the surrounding stand. Within each spatially explicit plot we recorded location (xy coordinate), diameter at breast height (dbh), and species for all trees >10 cm in dbh. Additionally, for each tree >10 cm in dbh we measured height using a clinometer. For the purpose of current comparison, a second, matched set of 1142 m2 systematic regeneration plots was established in the reserve stand in 2009. Regeneration and grazing dynamics plots were surveyed in 2009 with species and height class recorded for all seedlings.

2.2.2. Diversity We calculated diversity and evenness for both stands using species richness, Shannon index (H0 ), and Simpson’s index (D). Where pi is the proportion of individuals of the ith species when calculated using stem density and RBA when using basal area as the abundance measure, S is species richness, ni is the number of individuals in the ith species and N is the total number of individuals (Magurran, 2004). Shannon diversity (H0 ) where

X

2.2. Data analysis

H0 ¼ 

2.2.1. Community structure and composition We calculated summary statistics for each stand using Microsoft excel. We examined species composition, stem density in stems per hectare (sph), diameter distribution, and basal area (BA). We calculated basal area per hectare (BA/ha) and Relative Basal Area (RBA) for each species. We pooled diameters for plants with multiple stems to calculate total BA per individual. We analyzed height data using Analysis of Variance (ANOVA) and used Tukey’s adjusted pairwise comparisons to establish patterns of canopy stratification. We calculated importance value (IV) for each species, where IV = relative dominance (RBA) + relative density, because this is census data relative frequency was not included (Curtis and McIntosh, 1951; Skeen, 1973). For spatial analyses, and to establish patterns of canopy stratification, we combined the three similarly statured species of maple (Acer campbellii, Acer sterculiaceum, Acer hookerii), and then selected the nine most dominant species in the stand, based on RBA. We classified these species by evergreeness and relative canopy position to examine structural patterns. We used t-tests to test for differences in regeneration across stands (logged and reserve) (alpha = 0.05).

Shannon evenness measure (J0 ) where

pi ln pi

J 0 ¼ H= ln S Simpson Index (D) (expressed as 1/D) where

D¼

X ni ½ni  1 N½N  1

Simpson’s measure of Evenness (E) where

E¼

1=D S

2.3. Spatial pattern analysis For each of the nine most dominant species by basal area we performed spatial point pattern analysis. Basal area is a commonly used abundance measure in forested systems because it is closely related to total biomass, net primary production (NPP), carbon, and other commonly sought forest ecosystem metrics (Oliver and Larson, 1996; Barnes et al., 1998).

166

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

Numerous studies have successfully applied the Ripley’s K function to better understand the successional status of, and competitive relationships between species in forest stands, and the impact of spatial pattern on regeneration (Grau, 2000; Arevalo and Fernandez-Palacios et al., 2003; Salas et al., 2006; Franklin and Rey, 2007). We analyzed spatial patterns in the stand using the Ripley’s K function (Ripley, 2005):

KðsÞ ¼ k1 n1

X Iðdij < sÞ i–j

where k is the average density of the points, n is the number of points in the dataset, d is the Euclidian distance between ith and jth point of points in a dataset, and s is the spatial lag. We then plotted each based on the corrected L-function:

LðsÞ ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi KðsÞ=p

to determine whether the distributions of the species were completely spatially random (CSR), clumped, or dispersed using three commonly accepted edge corrections, border, translation, and isotropic (Ohser, 1983; Ripley, 1988; Waller and Gotway, 2004). We then applied a bi-variate Ripley’s K analysis:

K ij ðsÞ ¼ ðki kj AÞ1

XX k

l

Iðdik ;jl < sÞ

where A is the study area, and d is the distance between kth location of type i and the lth location of type j. We first tested each of the nine dominant species against each other. We then classified the individual trees into 4 groups based on calculated basal area, and canopy strata group as determined by Tukey’s pairwise comparison: (1) large old growth (all trees with a cumulative BA of more than .70 m2/ha, (2) canopy individuals (height > 20 m), (3) mid-story individuals (10 m > 20 m), understory individuals (<10 m). If an individual fell into more than one category BA took precedence over height in classification. We then performed bivariate spatial analyses between the large old growth trees and the three other classifications to test for the presence of treefall gaps driving canopy stratification. We also examined spatial relationships between deciduous and evergreen species. We used the isotropic edge correction for all bi-variate analyses. All calculations were performed in R, using the spatstat and splancs packages (Baddeley and Turner, 2005; R Development Core Team, 2010; Rowlingson and Diggle, 2012). Several species used in the spatial analysis had low stem counts in one or both stands (eg. B. alnoides). We chose to include these species because while relatively rare, individual stems of these species can be large, significantly affecting the spatial dynamics of the community. 3. Results 3.1. Species composition, and structure A total of 20 trees species were found across both plots. Three species were only present in the logged stand, Enkianthus deflexus, Fraxinus floribunda, and T. baccata; of those only E. deflexus had more than two individuals present. Stem density was higher in the reserve (435 sph compared to 384 sph), some of this difference is accounted for by reductions during experimental felling (Quercus and Tsuga). Density of shade-tolerant understory species (Corylus, Enkianthus, and Rhododendron) was higher in the logged stand. Snags were more numerous in the reserve; stumps more common in the harvested stand. The ten most abundant species by RBA throughout the two stands are Acer campbellii, A. hookeri, B. alnoides, Corylus ferox, Gamblea ciliata, Ilex dipyrena, Osmanthus suavis, Q. semecarpifolia, Rhododendron arboreum, and T. dumosa (Table 1). This assemblage consists of six evergreen and four deciduous species more or less evenly distributed in three distinct (p = 0) canopy strata (Table 1).

Diameter distributions for both stands show a reverse J-shape distribution, but the overall shape of the curve as well as individual species distributions are different between the two stands (Figs. 2 and 3). Following logging there is a distinct reduction in the number of large stems, particularly in canopy dominant T. dumosa and Q. semecarpifolia. These stems have largely been replaced by a dense understory of O. suavis, I. dipyrena, and R. arboreum. Accompanying this shift in species dominance is a notable reduction in the stature of this forest. 3.2. Regeneration The 114 systematic regeneration plots highlight the paucity of Quercus regeneration in both stands. Across both stands Q. semecarpifolia was found in only 6% of regeneration plots for a total number of seedlings equal to 5 or 4 seedlings per hectare in the logged and reserve stands respectively. Oak seedlings had an average height of 5–10 cm in both plots. Most notably, none of the Q. semecarpifolia seedlings measured in the original 2000 regeneration survey were still present during the 2009 sampling, implying that sparse oak regeneration is not establishing as longterm advance regeneration. At the time of this survey, no oak seedlings were present in any of the long-term grazing exclusion plots. T-tests showed no significant difference in the number of Quercus (p = .72), Tsuga (p = .39), or Acer (p = .12) seedlings between the harvested and reserved stands; the number of Enkianthus seedlings (p = .03) was significantly higher in the harvested stand whereas Rhododendron seedlings were more abundant in the reserve stand (p = .01) (Table 2). 3.3. Tree species diversity Tree species richness was higher in the logged stand (20 species versus 17). On the contrary, out of the six other diversity metrics calculated, only Shannon diversity (H0 ) calculated using stem counts as the abundance measure was similar between the logged and reserve stands, all other metrics pointed to lower diversity in the logged stand. H0 calculated using basal area as the abundance measure shows a larger discrepancy between the two stands. Evenness (J0 ) shows similar trends with the magnitude of difference in diversity between the logged and unlogged stands increasing when basal area is applied as the abundance measure. In fact, J0 is quite low (.23) in the logged stand compared to 0.47 in the unlogged stand. Evenness is important because it is both influenced by a different processes than species richness, and is often associated with different suite of environmental factors (Wilsey and Stirling, 2007). The Simpson diversity metrics also indicates lower diversity in the logged stands (Table 3). 3.4. Spatial pattern Stem density diagrams expose a lack of defined pattern in the logged stand, with little clumping or dispersion, whereas the reserve shows significant spatial clumping in areas lacking Q. semecarpifolia canopy gaps (Table 4). Analyses of spatial patterns using univariate Ripley’s K function for each of the nine most abundant species shows differences between the logged and reserve stands with most species trending towards spatial randomness (CSR) after logging. For example, in the logged stand Q. semecarpifolia, the most dominant species, displays CSR, whereas in the reserve stand the species shows marginal repulsion. Evergreen understory tree species R. arboreum and O. suavis were also randomly distributed in the logged stand. In the reserve stand these species were significantly clumped in gaps at scales greater than 15 m.

Table 1 All species with family, evergreen status, number of stems, importance value, and basal area for each stand. Relative canopy position and mean height is included for the ten most dominant species. Family

Sapindaceae Sapindaceae Sapindaceae Fabaceae Betulaceae Betulaceae Ericaceae Oleaceae Araliaceae Hydrangeaceae Aquifoliaceae Lauraceae Lauraceae Oleaceae Rosaceae Fagaceae Fagaceae Ericaceae Taxaceae Pinaceae

No No No No No No No No Yes No Yes No No Yes No Yes Yes Yes Yes Yes

Height (m) ± SD

13.1 ± 4.7 13.1 ± 4.7

Mid Mid

21.5 ± 6.1 9.5 ± 4.5

High Low

12.1 ± 3.0

Mid

11.7 ± 3.4

Mid

8.9 ± 3.5

Low

23.2 ± 10.8

High

8.7 ± 2.8

Low

32.4 ± 12.6

High

Total Live Stems Snags Stumps a b c

Canopy Positionc

Stand 1: Logged

Stand 2: Reserve

IV

BA (m2 ha1)

# Stems

IV

BA (m2 ha1)

# Stems

4.43 1.84 0.38 1.02 3.08 12.03 4.92 1.20 3.71 0.35 18.68 1.02 1.07 14.13 3.21 53.64 2.53 56.21 0.61 15.94

0.70 0.16 0.04 0.07 0.45 0.63 0.22 0.20 0.41 0.03 2.69 0.07 0.09 0.87 0.49 13.18 0.37 4.64 0.10 4.36

8 5 1 3 6 38 16 2 9 1 37 3 3 43 6 36 5 156 1 5

7.24 1.23 5.23 1.00 3.09 2.52 – – 2.23 0.80 15.22 2.94 0.60 15.20 0.50 91.24 1.33 27.51 – 22.12

1.35 0.05 2.49 0.06 0.70 0.15 – – 0.58 0.08 3.20 0.28 0.26 1.28 0.03 43.39 0.76 2.33 – 11.91

23 5 7 4 9 10 – – 6 3 46 11 1 58 2 123 1 105 – 21

29.77

384 15 12

68.91

435 26 2

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

Acer campbelliia,b Acer hookerii (syn. A. sikkimense)a,b Acer sterculiaceum Albizia odoratissima Betula alnoidesb Corylus feroxb Enkianthus deflexus Fraxinus floribunda Gamblea ciliatab Hydrangea heteromalla Ilex dipyrenab Lindera heterophylla Litsea kingii Osmanthus suavisb Prunus nepalensis Quercus semecarpifoliab Quercus thompsonii Rhododendron arboreumb Taxus baccata Tsuga dumosab

Evergreen

A. campbellii and A. hookeri combined into Acer spp. for spatial analyses. Ten most dominant species with additional information on canopy position used in spatial analysis. Relative canopy position grouping based on ANOVA and Tukey’s pairwise comparison p = 0.

167

168

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

Fig. 2. Diameter distributions for all live trees in both the logged and reserve stands.

Fig. 3. Diameter distributions for select dominant species in both the logged and reserve stands. (a) Q. semecarpifolia; (b) Acer spp.; (c) C. ferox; (d) I. dipyrena. Q. semecarpifolia was preferentially removed during experimental felling, resulting in far fewer total stems in the harvested stand, particularly in the larger size classes. A similar pattern can be observed in Acer spp., where as understory C. ferox and I. dipyrena show the opposite with far greater numbers of small stems after harvesting.

169

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173 Table 2 Summary of seedlings in the systematic regeneration plots in the logged and reserve stands. Total Seedlings/hectare

Acer Corylus Enkianthus Fraxinus Gamblea Hydrangea Ilex Lindera Litsea Picea Pieris Pinus Prunus Q. semecarpifolia Rhododendron Sorbus Tsuga Vaccinium

Stand 1: Logged

Stand 2: Reserve

2600 100 1550 50 200 400 1100 500 500 300 150 650 50 500 1300 550 1050 50

4250 150 150 0 50 0 750 400 100 0 0 0 0 400 1500 200 1800 250

Statistical Significance

Presence/Absence (%) Stand 1: Logged

Stand 2: Reserve

Stand 1: Logged

Stand 2: Reserve

p = .52 p = .70 p0 na p = .41 na p =. 74 p = .78 p = .12 na na na na p = .72 p = .01 p = .36 p = .79 p = .34

33 1 15 1 4 6 17 5 7 6 3 6 1 6 15 7 8 1

31 3 2 0 1 0 10 8 1 0 0 0 0 6 8 3 10 2

5–10 cm 11–25 cm 5–10 cm 11–25 cm 11–25 cm 11–25 cm 5–10 cm 5–10 cm 11–25 cm 5–10 cm 5–10 cm 5–10 cm 11–25 cm 5–10 cm 11–25 cm 26–50 cm 5–10 cm 11–25 cm

5–10 cm 5–10 cm 11–25 cm

Table 3 Summary of diversity measures for both the logged and unlogged stands. H0 c, Shannon diversity using stem count as the abundance measure; H0 BA, Shannon diversity using basal area as the abundance measure; J0 c, Shannon evenness stem count as abundance measure; J0 BA, Shannon evenness basal area as abundance measure; S, species richness; D, Simpson diversity expressed as 1/D; E, Simpson evenness.

H0 c H0 BA J0 c J0 BA S D E

Logged

Reserve

2.07 1.84 0.69 0.23 20 4.82 0.24

2.08 1.34 0.73 0.47 17 5.76 0.34

When all understory trees are grouped, we observe clumping at scales up to 30 m for both stands. The results in the logged stand indicate only minor effects, while in the reserve stand understory trees are clustered at all spatial scales. As a group, evergreens also show marginally significant clumping in the logged stand, significant clustering in the reserve. Considered together, deciduous species exhibit CSR in both stands. Bi-variate spatial analyses (Fig. 4) in the reserve stand reveal clumping between many species. For example, Betula (an early successional deciduous tree) and Rhododendron (an evergreen understory species) show clumping between species in the reserve, while in the logged stand these species exhibited CSR. The distribution of the dominant species Q. semecarpifolia to the understory evergreen G. ciliata displays CSR in the logged stand, but in the reserve there is dispersion between the two species at scales greater than 15 m. The two ecologically similar species G. ciliata and R. arboreum show CSR in the logged stand, but clumping even at close distances (>8 m) in the reserve. The relationship between Acer spp. with both Corylus and Osmanthus show the same trends; the logged stand shows CSR, the reserve shows clumping at higher distances. In the logged stand 75% of the pairings are CSR, while in the reserve we see only 33% (Table 4). 4. Discussion 4.1. Species composition, structure, and spatial pattern In general, the spatial attributes of the logged stand were more random and less heterogeneous than in the reserve stand (Table 4).

Mean Height Class

5–10 cm 5–10 cm 5–10 cm 5–10 cm

5–10 cm 5–10 cm 5–10 cm <5 cm <5 cm

This observation is in contrast to previous studies finding greater levels of spatial randomness in old growth trees and primary forests (Kuuluvainen et al., 1996; Youngblood et al., 2004; Salas et al., 2006). More heterogeneous structures likely lead to increased diversity as a greater number of unique microsites increase habitat (Rosenzweig, 1995; Báldi, 2008). In addition to its importance as habitat structure increased heterogeneity has been cited as an important driving factor in tree population dynamics and future spatial pattern formation (Getzin et al., 2008). This finding suggests the potential for these observed reductions in heterogeneity to persist. Higher densities of shade tolerant understory species (Rhododendron, Corylus, and Enkianthus), their progression into larger size classes, and significant clumping in the logged stand is likely the result of individuals being released following harvest. Noncommercial species are not commonly removed in rural harvest in Bhutan (RGOB, 2004). In the ten years since felling these noncommercial understory species have occupied released growing space, increasing their relative dominance. Past studies investigating spatial dynamics in old growth forests have found that young recruits tend to be centered in gaps and aggregated around dead trees (Moeur, 1993; Wolf, 2005; Salas et al., 2006). Indeed, we observed a similar pattern in the old growth reserve plot with significant spatial clumping among a variety of species and repulsion from canopy dominant Q. semecarpifolia and T. dumosa. Some species were found clumped throughout all gaps (Corylus), while other species were centered in only one or two large gaps (Acer, Betula, Gamblea, Rhododendron, and Tsuga). These individual species differences may be explained by stochastic events driving dispersal or could be the effect of interactions between species-specific tolerances and abiotic resource gradients (ex. light or moisture); however, the broader pattern of significant clumping in the reserve stand indicates that canopy dominants are dictating microsite habitat.

4.2. Regeneration The results of this study suggest that selective thinning in old growth Bhutanese evergreen oak stands may not be sufficient to regenerate Q. semecarpifolia. None of the seedlings tracked following selection harvesting in 2000 survived to 2009 and new oak seedling abundances were not significantly different from those in the old growth stand.

170

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

Table 4 Summary of univariate and bivariate Ripley’s K comparisons in both the logged and unlogged stands. Acer

Betula

Corylus

Gamblea

Ilex

Osmanthus

Logged Acer Betula Corylus Gamblea Ilex Osmanthus Quercus Rhododendron Tsuga

CSR CSR CSR CSR CSR CSR CSR Dispersed(>8 m) CSR

CSR CSR CSR CSR Clumpeda CSR CSR CSR

CSR CSR CSR CSR CSR CSR Clumped(>11 m)

CSR CSR CSR CSR CSR CSR

CSR CSR CSR CSR CSR

Clumped(<30) CSR CSR Clumpeda CSR CSR CSR

Reserve Acer Betula Corylus Gamblea Ilex Osmanthus

Clumped(>20 m) Dispersed(>20 m) Clumped(>8 m) Dispersed(>20 m) CSR Dispersed(>20 m)

Clumped(>15 m) Clumped(>15 m) Clumped(>15 m) CSR Clumped(>15 m)

Clumpeda Clumpeda Clumpeda Clumped

Clumped CSR Disperseda

Quercus CSR Rhododendron Disperseda Tsuga Clumpeda a

Disperseda Clumped(>0) CSR

Clumpeda Clumped (>15 m) CSR Dispersed(>15 m) CSR Clumped(>15 m) Clumped(>0) CSR Disperseda CSR CSR

Quercus

Rhododendron Tsuga

Clumped(>0) CSR

CSR

Clumpeda Disperseda Disperseda CSR

Disperseda Dispersed Clumped(>0) CSR CSR

Clumped(>20 m)

Marginally Significant.

Fig. 4. Graphical outputs of bivariate spatial relationships: (a) G. ciliata versus Q. semecarpifolia in the reserve stand showing significant dispersion (b) Acer spp. versus O. suavis in the reserve stand showing significant clumping (c) C. ferox versus G. ciliata in the reserve stand showing marginally significant clumping (d) Q. semecarpifolia versus R. arboreum in the harvested stand showing complete spatial randomness. The dashed centerlines represent the expected function of a random point pattern, shaded areas represent upper and lower confidence intervals, and solid lines are the observed function.

Wangda and Ohsawa (2006) reported ample regeneration in several undisturbed oak stands in Bhutan and proposed that small-scale human disturbance may be sufficient for regeneration. Other authors, however, have suggested that larger gaps (Metz, 1997) and the complete exclusion of grazing may be necessary (Davidson, 2000). Following individual seedlings over a 3-year

period after experimental felling Tashi and Thinley (2008) observed a correlation between canopy openness and seedling height growth, but not the number of seedlings. Furthermore, they noted increased vigor in Quercus seedlings inside fenced grazing exclosures. The authors observed a steep decline in the numbers of Quercus seedlings over the course of the study, but suggested

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

this could provide complete stocking if the few existing seedlings survived. Resurveying 10 years after harvest we found a complete absence of the few Quercus seedlings that established following harvest and very few new recruits in either the logged or reserve stands, suggesting dramatic declines in Quercus regeneration abundance. These findings suggest that Quercus has not successfully regenerated following single-tree harvesting. 4.3. Tree species diversity Tree species richness was moderately rich for a seasonal montane evergreen oak forest (17–20 species per hectare), when compared to similar systems. Tree species richness in 2.21 hectare belt transects in a similar system in southern Japan have 28 overstory species, consisting of sixteen evergreen broadleaf, eight deciduous broadleaf, and four coniferous species (Naka and Yoneda, 1984). Whereas Mediterranean evergreen oak forests have few canopy species consisting primarily of Quercus ilex, these forests are heavily managed and, their monospecificity is likely the result of historic human impacts (Rodà, 1999). While our census plots preclude us from statistically testing for differences in diversity between management histories, we can extrapolate some general trends. Our results suggest diversity, in general, is not much different between the logged and reserve stands; in fact, canopy tree richness may increase after timber harvesting. While a number of studies have shown positive effects on species richness following selective logging (Berry et al., 2008; Swaine and Agyeman, 2008; Su et al., 2010), others have shown decreases (Huang et al., 2003; Felton et al., 2006; Gradstein et al., 2007; Farwig et al., 2008), suggesting that factors other than harvesting may be influencing measures of diversity across these forest types (Clark and Covey, 2012). While it is difficult to be certain based on the available data, the increase in richness observed here is possibly the result of the release of shade-tolerant understory species (e.g. E. deflexus) moving into large size classes following release. Evenness, important because it is both influenced by a different processes than species richness, and is often associated with different suite of environmental factors (Wilsey and Stirling, 2007), on the other hand appears to be higher in the reserve stand regardless of the metric used. These results are in line with past work suggesting species evenness often is more sensitive than species richness to human activities and environmental change (Tilman, 1982; Hillebrand et al., 2008).

171

and Ohsawa, 2006). Well-timed overstory removals, however, may be needed to move established seedlings into the canopy (Metz, 1997; Tashi and Thinley, 2008). The timing and intensity of these release treatments should be tailored to mimic the nature and scale of the initiating disturbance. The current belief is that these stands emerged as the result of single-tree fall autogenic disturbance (Davidson, 2000). Ample evidence from stand reconstructions in other oak systems around the world suggests this is unlikely (Oliver, 1981; Suh and Lee, 1998; Harmer and Morgan, 2007; Dobrowolska, 2008). These studies, and many others, offer ample evidence that more intense disturbances favor the regeneration of oak species. The preference for colonizing after extreme disturbance is also evident in the body of research linking oak to abandoned pasture (Abrams, 1992; Thadani and Ashton, 1995; Quazi et al., 2003). The findings presented here are in keeping with the large body of research conducted in other oak systems, though further study could definitively address remaining questions, particularly those relating to the implications of grazing and the timing and scale of canopy disturbance. Experimental studies that combine grazing exclusion across canopy treatments from full opening to full shade, coupled with site preparation treatments such as scarification are warranted. Additionally, many of the studies from other systems rely heavily on age-class data gleaned from tree-ring data to reconstruct the timing and scale of stand initiating disturbance. Future work incorporating dendrochronological methods could provide a more complete reconstruction of stand history from which more complete silvicultural guidelines for this endangered system could be developed.

Acknowledgements The authors would like to acknowledge the important role the staff of the Ugen Wangchuck Institute for Conservation and the Environment, and the Renewable Natural Resources Research Centers at Yusipang, and Jakar played in providing technical and logistical support for this research. We would also like to thank Jonathan Reuning-Scherer and Tim Gregoire for providing statistical consultation and for reviewing early drafts of this work. This research was supported through grants from the Tropical Resources Institute, and the Jubitz Family fund at the Yale University School of Forestry and Environmental Studies. We would also like to acknowledge the thoughtful feedback from two anonymous reviewers.

5. Management implications Selective ungulate grazing is widely cited as limiting regeneration establishment (Belsky and Blumenthal, 1997; Vera, 2000; McEvoy et al., 2006). While intense grazing pressure almost certainly played a role in restricting regeneration to less palatable species, the absence of Quercus inside grazing exclosures indicates that cattle browse is likely not the sole cause of oak regeneration failure. As observed elsewhere, restricted gap size and a dense evergreen understory may be limiting Quercus growth following harvest (Barton et al., 1994; Lorimer et al., 1994; Moser et al., 1996). Indeed, balancing the need to suppress early successional competition with the light availability needed to drive oak seedling and sapling growth is regarded as the central challenge of oak silviculture (Johnson et al., 2002). Spatial analyses show clumping and repulsion for a number of species range from 15–30 m in the natural stand (Table 4), this may indicate that naturally occurring gaps in the reserve are at this scale. These gaps may be sufficient to allow for the establishment of advanced regeneration in Q. semecarpifolia, particularly if harvests remove competing under- and mid-story species (Wangda

References Abrams, M.D., 1992. Fire and the development of oak forests. Bioscience 42, 346– 353. Abrams, M.D., Orwig, D.A., Demeo, T.E., 1995. Dendroecological analysis of successional dynamics for a presettlement-origin white-pine mixed-oak forest in the southern appalachians, USA. J. Ecol. 83, 123–133. Arevalo, J.R., Fernandez-Palacios, J.M., 2003. Spatial patterns of trees and juveniles in a laurel forest of Tenerife. Canary Islands Plant Ecol. 165, 1–10. Baddeley, A., Turner, R., 2005. Spatstat: an R package for analyzing spatial point patterns. J. Statist. Software, 1–42. Baker, P.J., Bunyavejchewin, S., Oliver, C.D., Ashton, P.S., 2005. Disturbance history and historical stand dynamics of a seasonal tropical forest in western Thailand. Ecol. Monogr. 75, 317–343. Báldi, A., 2008. Habitat heterogeneity overrides the species–area relationship. J. Biogeogr. 35, 675–681. Barnes, B.V., Zak, D.R., Denton, S.R., Spurr, S.H., 1998. Forest Ecology. John Wiley & Sons Inc, New York, NY. Barton, D.C., Boring, L.R., Swank, W.T., 1994. Regeneration patterns in canopy gaps of mixed-oak forests of the Southern Appalachians: influences of topographic position and evergreen understory. Am. Midl. Nat. 132, 308–319. Bean, W.J., 1916. Trees and shrubs hardy in the British Isles/by W.J. Bean. J. Murray, London. Belsky, A.J., Blumenthal, D.M., 1997. Effects of livestock grazing on stand dynamics and soils in upland forests of the interior west. Conserv. Biol. 11, 315–327.

172

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173

Berry, N.J., Phillips, O.L., Ong, R.C., Hamer, K.C., 2008. Impacts of selective logging on tree diversity across a rainforest landscape: the importance of spatial scale. Landscape Ecol. 23, 915–929. Bisht, H., Prakash, V., Nautiyal, A., 2012. Factors Affecting Regeneration Potential of Quercus semecarpifolia, Smith: A Poor Regenerated Oak of Himalayan Timberline. Res. J. Seed Sci., 5. Biswas, A.K., 1986. Forestry and forest management in Bhutan. Resources Policy 12, 145–148. BSS/NSSC, 2003. Soil report on forest regerneration trial site, Chimithanka, Thimphu. In: Service, B.S. (Ed.). National Soil Survey Center, Chimithanka Thimphu. Buffum, B., Gratzer, G., Tenzin, Y., 2008. The sustainability of selection cutting in a late successional broadleaved community forest in Bhutan. For. Ecol. Manage. 256, 2084–2091. Clark, J.A., Covey, K.R., 2012. Tree species richness and the logging of natural forests: A meta-analysis. For. Ecol. Manage. 276, 146–153. Curtis, J.T., McIntosh, R.P., 1951. An Upland Forest Continuum in the Prairie-Forest Border Region of Wisconsin. Ecology 32, 476–496. Davidson, J., 2000. Ecology and Management of the Broadleaved Forests of Eastern Bhutan. In: MOA, D.o.F.S. (Ed.). Royal Government of Bhutan, Thimphu. Dhar, U., Rawal, R.S., Samant, S.S., 1997. Structural diversity and representativeness of forest vegetation in a protected area of Kumaun Himalaya, India: implications for conservation. Biodivers Conserv. 6, 1045–1062. Dobrowolska, D., 2008. Effect of stand density on oak regeneration in flood plain forests in Lower Silesia, Poland. Forestry 81, 511–523. Farwig, N., Braun, C., Boehning-Gaese, K., 2008. Human disturbance reduces genetic diversity of an endangered tropical tree, Prunus africana (Rosaceae). Conserv. Genet. 9, 317–326. Felton, A., Felton, A.M., Wood, J., Lindenmayer, D.B., 2006. Vegetation structure, phenology, and regeneration in the natural and anthropogenic tree-fall gaps of a reduced-impact logged subtropical Bolivian forest. For. Ecol. Manage. 235, 186– 193. Fischer, F., 1976. Bhutan: the importance of the forests for a continuous development of human ecology in high mountain conditions. In. Bhutan: the importance of the forests for a continuous development of human ecology in high mountain conditions. [Swiss Federal Institute of Technology]., Zurich, Switzerland, vol. 158, 1976, p. 31. Franklin, J., Rey, S.J., 2007. Spatial patterns of tropical forest trees in Western Polynesia suggest recruitment limitations during secondary. J. Trop. Ecol. 23, 1– 12. Gamble, J.S., 1881. A Manual of Indian Timbers: an Account of the Structure, Growth, Distribution, and Qualities of Indian Woods. Office of the Superintendent of Govt. Printing, Calcutta. Getzin, S., Wiegand, T., Wiegand, K., He, F., 2008. Heterogeneity influences spatial patterns and demographics in forest stands. J. Ecol. 96, 807–820. Gradstein, S., Kessler, M., Pitopang, R., 2007. Tree species diversity relative to human land uses in tropical rain forest margins in Central Sulawesi. Stability of Tropical Rainforest Margins: Linking Ecological, Economic and Social Constraints of Land Use and Conservation, pp. 321–334. Grau, H.R., 2000. Regeneration patterns of Cedrela lilloi (Meliaceae) in northwestern Argentina subtropical montane forests. J. Trop. Ecol. 16, 227–242. Harmer, R., Morgan, G., 2007. Development of Quercus robur advance regeneration following canopy reduction in an oak woodland. Forestry 80, 137–149. Hillebrand, H., Bennett, D.M., Cadotte, M.W., 2008. Consequences of dominance: a review of evenness effects on local and regional ecosystem processes. Ecology 89, 1510–1520. Huang, W., Pohjonen, V., Johansson, S., Nashanda, M., Katigula, M., Luukkanen, O., 2003. Species diversity, forest structure and species composition in Tanzanian tropical forests. Forest Ecol. Manag. 173, 11–24. Huxley, A.J., Griffiths, M., Levy, M., Society, R.H., 1992. The New Royal Horticultural Society Dictionary of Gardening. Macmillan Press. Jackson, J.K., 1984. Manual of Afforestation in Nepal. Forest Research and Survey Centre. Johnson, P.S., Shifley, S.R., Roberts, R., 2002. The Ecology and Silviculture of Oaks. CABI, New York, NY. Kuuluvainen, T., Antti, P., Keinonen, K., Markku, N., 1996. Statistical opportunities for comparing stand structural heterogeneity in managed and primeval forests: an example from boreal spruce forest in southern Finland. Silva Fenn. 30, 315– 328. Liptzin, D., Ashton, P.M.S., 1999. Early-successional dynamics of single-aged mixed hardwood stands in a southern New England forest, USA. For. Ecol. Manage. 116, 141–150. Lorimer, C.G., Chapman, J.W., Lambert, W.D., 1994. Tall understorey vegetation as a factor in the poor development of oak seedlings beneath mature stands. J. Ecol. 82, 227–237. Magurran, A.E., 2004. Measuring Biological Diversity. Blackwell Publishing Company, Malden, MA. McEvoy, P.M., McAdam, J.H., Mosquera-Losada, M., Rigueiro-Rodriguez, A., 2006. Tree regeneration and sapling damage of pedunculate oak Quercus robur in a grazed forest in Galicia, NW Spain: a comparison of continuous and rotational grazing systems. Agroforest Syst. 66, 85–92. Metz, J.J., 1997. Vegetation dynamics of several little disturbed temperate forests in East Central Nepal. Mt. Res. Dev. 17, 333–351. Moeur, M., 1993. Characterizing spatial patterns of trees using stem-mapped data. Forest Sci. 39, 756–775.

Moser, W.K., Ducey, M.J., Ashton, P.M.S., 1996. Effects of fire intensity on competitive dynamics between red and black oaks and mountain laurel. Northern J. Appl. Forestry 13, 119–123. Naka, K., Yoneda, T., 1984. Community dynamics of evergreen broadleaf forests in southwestern Japan. The botanical magazine = Shokubutsu-gaku-zasshi vol. 97, pp. 275–286. Ohser, J., 1983. On estimators for the reduced second moment measure of point processes. Mathematische operationsforschung und statistik, Series Statistics 14, 63–71. Oliver, C.D., 1981. Forest development in North-America following major disturbances. For. Ecol. Manage. 3, 153–168. Oliver, C.D., 1992. Similarities of Stand Structure Patterns Based on Uniformities of stand Development Processes Throughout the World–Some Evidence and the Application to Silviculture Through Adaptive Management. In: Kelty, M.J., Larson, B.C., Oliver, C.D. (Eds.), The Ecology and Silviculture of Mixed-Species Forests: A Festschrift for David M. Smith. Kluwer Academic Publishers, Boston, pp. 11–26. Oliver, C.D., Larson, B.C., 1996. Forest Stand Dynamics. John Wiley & Sons Inc. Oliver, C.D., Stephens, E.P., 1977. Reconstruction of a Mixed-Species Forest in Central New England. Ecology 58, 562–572. Orwa, C., Mutua, A., Kindt, R., Jamnadass, R., Anthony, S., 2009. Agroforestree Database: a tree reference and selection guide version 4.0. In. Quazi, S.A., Ashton, M.S., Thadani, R., 2003. Regeneration of Monodominant Stands of Banj Oak (Quercus leucotrichophora A. Camus) on Abandoned Terraces in the Central Himalayas. J. Sustain. Forestry 17, 75–90. R Development Core Team, 2010. R: A Language and Environment for Statistical Computing. In: R Foundation for Statistical Computing, Vienna, Austria. Reif, A., Jolitz, T., Munch, D., Bucking, W., 1999. Succession from a Quercus-Carpinus forest towards an Acer forest: Process of natural tree species regeneration in the forest reserve ‘Bechtaler Wald’ near Kenzingen, southwest Germany. Allg. Forst Jagdztg. 170, 67–74. RGOB, 2004. Forest Management Code of Bhutan, Thimphu. Ripley, B.D., 1988. Statistical Inference for Spatial Processes. Cambridge University Press, New York. Ripley, B.D., 2005. Spatial statistics. John Wiley & Sons. Rodà, F., 1999. Ecology of Mediterranean Evergreen Oak Forests. Springer. Rosenzweig, M.L., 1995. Species Diversity in Space and Time. Cambridge University Press. Rowlingson, B., Diggle, P., 2012. splancs, spatial and space-time point pattern analysis package. In. Ruffner, C.M., Abrams, M.D., 1998. Relating land-use history and climate to the dendroecology of a 326-year-old Quercus prinus talus slope forest. Can. J. Forest Research-Revue Canadienne De Recherche Forestiere 28, 347–358. Salas, C., Lemay, V., Nunez, P., Pacheco, P., Espinosa, A., 2006. Spatial patterns in an old-growth Nothofagus obliqua forest in south-central Chile. For. Ecol. Manage. 231, 38–46. Sano, J., 1997. Age and size distribution in a long-term forest dynamics. For. Ecol. Manage. 92, 39–44. Sargent, C., 1985. The forests of Bhutan. Ambio 14, 75–80. Shrestha, B.B., 2003. Quercus semecarpifolia Sm. in the Himalayan region: Ecology, exploitation and threats. J. Himalyan Sci. 1, 126–128. Singh, S., Singh, J., 1986. Structure and function of the Central Himalayan oak forests. Proceedings: Plant Sciences vol. 96, pp. 159–189. Skeen, J.N., 1973. An extension of the concept of importance value in analyzing forest communities. Ecology 54, 655–656. Smith, D.M., Larson, B.C., Kelty, M.J., Ashton, P.M.S., 1997. The Practice of Silviculture Applied Forest Ecology. John Wiley and Sons, INC., Hoboken, NJ. Su, D., Yu, D., Zhou, L., Xie, X., Liu, Z., Dai, L., 2010. Differences in the structure, species composition and diversity of primary and harvested forests on Changbai Mountain, Northeast China. J. Forest Sci. (Prague) 56, 285–293. Suh, M.H., Lee, D.K., 1998. Stand structure and regeneration of Quercus mongolica forests in Korea. For. Ecol. Manage. 106, 27–34. Swaine, M.D., Agyeman, V.K., 2008. Enhanced Tree Recruitment Following Logging in Two Forest Reserves in Ghana. Biotropica 40, 370–374. Tashi, S., Thinley, C., 2008. Regeneration of Brown Oak (Quercus semecarpifolia Sm.) in an Old Growth Oak Forest. J. Renew. Nat. Resources Bhutan 4, 11–24. Thadani, R., Ashton, P.M.S., 1995. Regeneration of Banj Oak J (QuercusLeucotrichophora Camus, A.) in the Central Himalaya. For. Ecol. Manage. 78, 217–224. Tilman, D., 1982. Resource Competition and Community Structure. Princeton University Press. Troup, R.S., 1921. The Silviculture of Indian Trees. Clarendos Press, Oxford, U.K.. Upreti, N., Tewari, J.C., Singh, S.P., 1985. The Oak Forests of the Kumaun Himalaya (India): Composition, Diversity, and Regeneration. Mt Res Dev 5, 163–174. Vera, F.W.M., 2000. Grazing Ecology and Forest History. CABI Publishing, New York. Verma, A., Tewari, A., Shah, S., 2012. Carbon storage capacity of high altitude Quercus semecarpifolia, forests of Central Himalayan region. Scand. J. Forest Res. 27, 609–618. Vetaas, O.R., 2000. The Effect of Environmental Factors on the Regeneration of Quercus semecarpifolia Sm. in Central Himalaya. Nepal Plant Ecol. 146, 137–144. Waller, L.A., Gotway, C.A., 2004. Applied Spatial Statistics for Public Heath Data. John Wiley and Sons Inc., Hoboken, NJ. 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. For. Ecol. Manage. 230, 136–150.

K. Covey et al. / Forest Ecology and Management 336 (2015) 163–173 Wilsey, B., Stirling, G., 2007. Species richness and evenness respond in a different manner to propagule density in developing prairie microcosm communities. Plant Ecol. 190, 259–273. Wolf, A., 2005. Fifty year record of change in tree spatial patterns within a mixed deciduous forest. For. Ecol. Manage. 215, 212–223.

173

Youngblood, A., Max, T., Coe, K., 2004. Stand structure in eastside old-growth ponderosa pine forests of Oregon and northern California. For. Ecol. Manage. 199, 191–217.