Anthropogenic disturbance and tree diversity in Montane Rain Forests in Chiapas, Mexico

Anthropogenic disturbance and tree diversity in Montane Rain Forests in Chiapas, Mexico

Forest Ecology and Management 154 (2001) 311±326 Anthropogenic disturbance and tree diversity in Montane Rain Forests in Chiapas, Mexico NeptalõÂ Ram...

327KB Sizes 0 Downloads 14 Views

Forest Ecology and Management 154 (2001) 311±326

Anthropogenic disturbance and tree diversity in Montane Rain Forests in Chiapas, Mexico NeptalõÂ RamõÂrez-Marciala,b,*, Mario GonzaÂlez-Espinosab, Guadalupe Williams-Linerac a

Posgrado en EcologõÂa y Manejo de Recursos Naturales, Instituto de EcologõÂa, A.C., Apartado Postal 63, C.P. 91000, Xalapa, Veracruz, Mexico b Departamento de EcologõÂa y SistemaÂtica Terrestres, DivisioÂn de ConservacioÂn de la Biodiversidad, El Colegio de la Frontera Sur (ECOSUR), Apartado Postal 63, C.P. 29200, San CristoÂbal de Las Casas, Chiapas, Mexico c Departamento de EcologõÂa Vegetal, Instituto de EcologõÂa, A.C., Apartado Postal 63, C.P. 91000, Xalapa, Veracruz, Mexico Received 8 August 2000; accepted 30 October 2000

Abstract We studied the in¯uence of anthropogenic disturbance on forest structure and composition in the highly populated Montane Rain Forests of northern Chiapas, Mexico. We evaluated species richness, basal area and stem density on 81 circular plots (0.1 ha each) along a categorical disturbance gradient due to forest extraction, livestock grazing, and ®res. A total of 116 tree species (>5 cm DBH) were recorded in three major forest types recognized by TWINSPAN. The three forest types were: Quercus±Podocarpus Forest (QPF), Pinus±Quercus±Liquidambar Forest (PQLF), and Pinus Forest (PF). The number of canopy and understory trees species, absolute density, and basal area decreased with disturbance intensity. Mean basal area of Pinus spp. was high at intermediate and severe disturbed sites (27 and 19 m2 ha 1, respectively), and low (0.2 m2 ha 1) in well preserved old-growth stands. Distribution of life forms was heterogeneous among forest types, with a high number of understory trees species in QPF, and an impoverished composition in PF. A ®rst axis obtained by factor analysis, represented a combination of anthropogenic disturbance along with environmental and structural variables. Scores of the ®rst factor explained almost 50% of variation, and was positively correlated with livestock grazing, ®rewood extraction, basal area of Pinus spp. and soil pH, and negatively associated with elevation, plant cover and basal area of Quercus spp. A second factor explained an additional 12% of variation and was associated with forest ®res and timber extraction. Distribution of size classes in the QPF was signi®cantly different …p < 0:05† than in the other two forest types, including the largest individuals in all inventories. Our results suggest that small scale, but frequent anthropogenic disturbance, increases the dominance of Pinus and drastically decreases ¯oristic richness, mostly understory trees. This points to the need of developing restoration practices aimed to attain highly diverse mixed forests from induced depauperate pinelands. On the other hand, the remnant MRF stands are currently under risk of deforestation in a highly populated Mayan territory, and their conservation under criteria of sustainable use may require ®nding alternative high value uses not included in conventional commercial forestry. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Floristic diversity; Forest fragmentation; Land-use; Pine±oak±sweetgum forest; Succession

*

Corresponding author. Present address: El Colegio de la Frontera Sur (ECOSUR), Apartado Postal 63, C.P. 29200, San CristoÂbal de Las Casas, Chiapas, Mexico. Tel.: ‡52-967-81883, ext.: 5104; fax: ‡52-967-82322. E-mail address: [email protected] (N. RamõÂrez-Marcial). 0378-1127/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 2 7 ( 0 0 ) 0 0 6 3 9 - 3

312

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

1. Introduction Natural disturbances determine forest dynamics and tree diversity at the local and regional scales (Pickett and White, 1985; Clark, 1992; Young and Hubbell, 1991; Attiwill, 1994; Burslem and Whitmore, 1999; Hubbell et al., 1999; Masaki et al., 1999; Sheil, 1999). Anthropogenic disturbance may contribute to regulate the regeneration dynamics, structure and ¯oristic composition of tropical lowland and Montane Rain Forests (Ewel et al., 1981; GoÂmez-Pompa and VaÂzquezYanes, 1981; Horn and Hickey, 1991; Hong et al., 1995). Relatively minor changes in structural attributes of forests have been related to disturbance regimes involving mostly a single factor that is limited in extent (e.g. selective logging; Cannon et al., 1994; Vetaas, 1997; Nagaike et al., 1999; Scherer et al., 2000). However, frequent, but low-intensity disturbance (e.g. grazing and browsing, ®rewood extraction) may involve the combined effect of multiple factors, and may strongly affect forest structure and the ability of understory species to regenerate in the disturbed area (Veblen, 1978; GonzaÂlez-Espinosa et al., 1991, 1995a; Smiet, 1992; Kappelle et al., 1996). The Montane Rain Forest (MRF; sensu Breedlove, 1981), sometimes referred to as Tropical Montane Cloud Forest (StadtmuÈller, 1987; Hamilton et al., 1995), or Bosque Meso®lo de MontanÄa (Rzedowski, 1978), is widely considered as one of the most diverse neotropical formations (Webster, 1995; Churchill et al., 1995; Rzedowski, 1996; Aldrich et al., 1997). Stands of MRF in the northern highlands of Chiapas (SE Mexico) have received intensive human disturbance for centuries, mainly in relation to slash-andburn agriculture (milpa system) practiced by Mayan and mestizo people (Collier, 1975). More recently, large areas of MRF (>40% of its original distribution; Challenger, 1998) have been fragmented or permanently cleared for extensive raising of sheep and cattle (grazing may occur also in closed forested habitats). Other human activities in the region include intensive selective logging for fuelwood and timber, and the establishment of coffee plantations (LoÂpez-Carmona, 1999). As a result of this landuse pattern, many primary MRF fragments persist intermingled with secondary pine±oak forests, pinelands, coppices, and shrublands (Zuill and Lathrop, 1975; Breedlove, 1981; Morales-HernaÂndez, 2000;

see Ochoa-Gaona and GonzaÂlez-Espinosa (2000) for a detailed treatment of a similar pattern in the central highlands of Chiapas). Population increases of Pinus spp. in many regions of the world, both as an exotic or native group, have been related to anthropogenic factors (Richardson and Bond, 1991; GonzaÂlez-Espinosa et al., 1995a,b; Schneider, 1996; Savage, 1997; Richardson, 1998; Stapanian and Cassell, 1999). Selective logging of coexisting canopy components (e.g. Quercus spp.) may favor local increases of pines. The resulting pinelands are less diverse in woody elements, their soils are somewhat less fertile (signi®cantly lower cation exchange capacity, and lower content of nitrogen and organic matter), and show more extreme microclimatic conditions (Zuill and Lathrop, 1975; Camacho-Cruz et al., in press; L. Galindo-Jaimes et al., unpublished). The objectives of this study are: (1) to determine the effect of human disturbance on dominance relations of canopy tree species and (2) to determine if changes in canopy dominance by Pinus and Quercus within plant associations contribute to account for richness of understory tree species. 2. Methods 2.1. Study sites and land-use history The study was conducted in 81 stands located in six localities along the crests and SW slopes of ranges at the municipality of Pueblo Nuevo SolistahuacaÂn in the northern highlands of Chiapas, southeastern Mexico (178080 ±178150 N; 928520 ±938000 W; Table 1). The stands studied include a wide range of successional stages, from relatively undisturbed old-growth forests, corresponding to one of last large remnants of MRFs, to heavily disturbed and open stands. The area encompasses ca. 740 km2, it is characterized by steep slopes (30±608, but more commonly 20±458), and elevation ranges from 1700 up to 2300 m. Karst formations prevent the occurrence of permanent rivers in the region, and only small creeks increase their ¯ow during the rainy season. Annual rainfall is 1400±2000 mm, and heavy fog occurs for 4±8 h every day during 8±10 months in the highest locations (Miranda, 1952; Breedlove, 1981; Zuill and Lathrop, 1975).

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

313

Table 1 Characteristics of study sites at the municipality of Pueblo Nuevo SolistahuacaÂn, northern highlands of Chiapas, Mexico Localities

No. of plots Elevation (m) Latitude (N) Longitude (W) Slope angle (8) Soil pH Predominant forest typea Canopy cover (%) a

RincoÂn Chamula

Cerro Blanco

Colegio Lindavista

El Campanario

La Yerbabuena

Los Lotes

19 1680±1810 178110 928550 4±60 5.9±6.7 PF 17±60

6 2200±2300 178140 928550 17±39 5.2±5.7 QPF 53±90

15 1780±1900 178110 928540 20±45 5.5±6.2 PQLF 40±80

2 1700±1710 178080 928530 18±22 6.4±6.7 PQLF 45±75

29 2080±2200 178100 928530 3±60 5.1±6.1 PQLF, QPF 41±90

10 1670±1810 178080 928520 3±28 5.5±6.5 PQLF 50±90

PF: Pinus Forest; PQLF: Pinus±Quercus±Liquidambar Forest; QPF: Quercus±Podocarpus Forest.

Notwithstanding that density of human population is relatively smaller than in other mountainous regions of Chiapas (Morales-HernaÂndez, 2000; Ochoa-Gaona and GonzaÂlez-Espinosa, 2000), most of the mature forest has been converted into milpa ®elds under slashand-burn system. The remaining secondary forests are subjected to ®rewood and timber extraction for subsistence and commercial purposes, and more recently to extensive livestock grazing in hillsides (LoÂpezCarmona, 1999). As a result of the neo-Zapatista indigenous Mayan rebellion of 1994 (Collier, 1994), ca. 95% of private lands changed to communal land tenure system; the new owners intend to operate a plan for sustainable forest management in the so-called ``recovered Indian territories'' (GonzaÂlez-HernaÂndez and Quintanar, 1999).

and botanical descriptions (Standley et al., 1946±1976). We classi®ed as canopy trees those species having the top of their crowns above any other tree layers, and fully exposed to sunlight in late successional stands. Understory trees were those species with stems

2.2. Floristics and vegetation structure Floristic composition (total number of woody species) and stand structure (density and basal area of stems in several diametric classes; Fig. 1) were evaluated in each of 81 circular plots (0.1 ha each) between September 1997 and March 1999. All individuals were recorded according to size categories in nested circular plots following a protocol for inventory of permanent forest plots in Mexico (after OlveraVargas et al., 1996; Fig. 1). We evaluated at least six plots in each locality, except at El Campanario where only two plots could be measured. Woody stems and seedlings were counted and identi®ed, and life form (either shrub, understory tree, or canopy tree) was de®ned for each species based on ®eld observations

Fig. 1. Shape and size details of the plots used for forest inventory in northern highlands of Chiapas, Mexico (adapted from OlveraVargas et al., 1996). A: seedlings (<0.5 m height) in 4  2 m2 plot (0.8 m of plot radius); B: saplings and juveniles (>0.5 m tall and <5 cm DBH) in 4  8 m2 plot (1.6 m of plot radius); C: small sized trees (5±10 cm DBH) in 1  100 m2 plot (5.6 m of plot radius); D: mid-sized trees (10±30 cm DBH) in 1  500 m2 plot (12.6 m of plot radius); E: large trees (>30 cm DBH) in 1  1000 m2 plot (17.8 m of plot radius).

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

314

>5 cm DBH and >4 m tall, whose crowns are usually underneath canopy species in well-developed stands. We classi®ed as shrubs those species whose adults had several branches at their bases, a maximum height of 4 m and DBH < 10 cm. 2.3. Environmental and anthropogenic disturbance variables Elevation, slope angle, and soil pH (measured at three random points with a KelwayTM portable soil pH-meter) were obtained for each plot. Forest cover was estimated with a convex densiometer (Forestry SuppliersTM) at four ®xed positions within the 500 m2 subplot. Recent evidence of forest extraction (at least 10 years before), frequency and intensity of livestock grazing, and ®re were evaluated. Forest extraction included separate estimates for ®rewood (number of stumps <20 cm DBH), timber (number of stumps >20 cm DBH), and resin (number of Pinus spp. individuals with evident collecting cuts on their bark; ocoteo is the common name of the practice). Grazing/browsing by livestock was assigned a categorical value, based on observations of stock density, abundance of feces, signs of browsing, and number of years since grazing occurred in each plot. Fire frequency and intensity in the plots was estimated using the number of burnt stumps, presence of ashes

on the forest ¯oor, and number of years since last ®re (see Table 2). 2.4. Data analysis The TWINSPAN program (Hill, 1979) was used to classify the stands. A matrix of 116 species (with stems >5 cm DBH) and 81 plots was built using relative importance values (relative density and relative dominance) for each species. The standard options of the program were used. The ®rst two divisions represented the major forest types in the region; a third division was obtained to segregate major plant associations within forest types. Species richness, mean density, and mean basal area (using the natural logarithm transformation) were analyzed for diametric categories with one-way ANOVA within groups obtained by TWINSPAN. Frequency distributions of size classes (DBH) among forest types were pairwise compared with Kolmogorov±Smirnov two-sample tests (Siegel, 1956). The coef®cient of variation of size classes was calculated for the most abundant species within each forest type (preferred over Gini's coef®cient of inequality; see Bendel et al., 1989), and pairwise compared with Mann±Whitney U-tests (Siegel, 1956). Factor analysis (Dillon and Goldstein, 1984) was used to identify underlying factors that could explain

Table 2 Categories of anthropogenic disturbance variables recorded in the MRF in the northern highlands of Chiapas, Mexico Source of anthropogenic disturbance

Scale of intensity Absent (0)

Low (1)

Moderate (2)

Severe (3)

Forest extraction Firewooda Timberb Resinc

None None None

1±40 1±20 1±2

41±100 21±50 3±5

>100 >50 >5

Livestock grazing Presence of feces Browsing trace Frequency of occurrenced

None None None

Occasional Occasional <3

Frequent Frequent 3±10

Heavily Heavily >10

Forest fires Burned stumps, presence of ashes No. of years since last fire

None None

Occasional 1±2

Frequent 3±10

Heavily >10

a

No. of stumps <20 cm DBH (ind. ha 1). No. of stumps >20 cm DBH (ind. ha 1). c No. of peeled out stems of Pinus (ind. ha 1). d Records of occurrence in years. b

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

relations among a set of environmental, anthropogenic, and structural variables. All variables were standardized by the mean normal variant before analysis. We used the maximum-likelihood method for factor extraction. Interpretation of the scores was attained evaluating the factor loading (eigenvectors) for each variable. Linear regression was used to explain the number of species by life form (canopy trees, understory trees, or shrubs) as a dependent variable in relation to the ®rst factor as explanatory variable (see LoÂpez-Portillo and Ezcurra (1989) for a similar treatment of data). All statistical analysis was undertaken using SPSS, v. 8.0 (Anonymous, 1998). 3. Results 3.1. Floristic composition and forest structure A total of 179 woody species of 116 genera belonging to 65 families were recorded. We recognized 23 species (13%) as canopy trees, 93 species (52%) as understory trees, and 64 species (35%) as shrub species, including four shrubby ferns and palms. The most specious families were Asteraceae (15 species), Leguminosae (11), Rubiaceae (11), Fagaceae (9), Lauraceae (9), Myrsinaceae (7), Araliaceae (6), and Aquifoliaceae (5).

315

After three divisions TWINSPAN suggested seven plant associations included within three major forest types for all stems >5 cm DBH (Fig. 2). The Quercus± Podocarpus Forest (QPF) was dominated by Q. benthamii NeÂe and P. matudai Lundell, and included three associations with 65% of the total number of species in the study area. Associations also included Persea schiedeana Nees and Calyptranthes pallens Lundell (1a), Clethra macrophylla Mart. & Gal. (1b), Turpinia tricornuta Lundell, and Magnolia sharpii Miranda (1c). The Pinus±Quercus±Liquidambar Forest (PQLF) accounted for 83% of the total ¯oristic richness, including Pinus tecunumanii (MartõÂnez) Eguiluz & Perry, Quercus candicans NeÂe, and sweetgum, Liquidambar styraci¯ua L. as canopy dominants. Two associations within this forest type included Nyssa sylvatica Marsh (2a) and Cleyera theaeoides (Sw.) Choisy (2b). The Pinus Forest (PF) represented 29% of the total richness and had Pinus oocarpa Schiede ex Schldl. as dominant species. The understory layer was typically species-poor, and included mostly secondary trees such as Acacia pennatula (Schldl. & Cham.) Benth., Myrica cerifera L., and Vernonia canescens H.B.K. Two associations (3a and 3b) also included Arbutus xalapensis H.B.K. and Pinus maximinoi H.E. Moore, respectively. Physiognomy and structure were similar for stands within associations, but differed among the three

Fig. 2. TWINSPAN divisions for the MRF communities in the northern highlands of Chiapas, Mexico. Eigenvalues between brackets. Distinguished communities are: (1) Q. benthamii±P. matudai Forest (QPF), subdivided into (1a) an association with P. schiedeana and Callyptranthes pallens, (1b) an association with C. macrophylla, and (1c) an association with T. tricornuta and M. sharpii; (2) P. tecunumanii± Q. candicans±L. styraci¯ua Forest (PQLF), subdivided into (2a) an association with N. sylvatica and (2b) an association with C. theaeoides; (3) P. oocarpa Forest (PF), subdivided into (3a) an association with A. xalapensis, and (3b) an association with P. maximinoi±A. pennatula.

316 Table 3 Mean species richness, density, and basal area (standard error in parentheses) for seven plant associations into the three forest types in the northern highlands of Chiapas, Mexicoa QPF

PQLF

PF

Total {81}

1b {7}

1c {3}

2a {12}

2b {15}

3a {15}

3b {4}

Species richness Canopy trees in 0.1 ha Understory trees in 0.1 ha Shrubs in 500 m2 Large trees in 0.1 ha Mid-sized trees in 500 m2 Small trees in 100 m2 Saplings in 8 m2 Seedlings in 2 m2

4 (0.2) a 19 (0.7) a 8 (0.3) ab 8 (0.5) a 11 (0.6) a 5 (0.7) a 19 (1.0) a 16 (0.9) bc

4 (0.6) a 23 (1.3) b 11 (1.2) a 7 (0.9) ab 14 (1.4) b 5 (1.1) ab 21 (2.4) a 21 (2.2) a

3 (1.2) a 23 (0.9) ab 8 (1.5) ab 6 (0.6) ab 9 (1.8) abc 6 (1.4) a 22 (0.8) a 19 (2.3) abc

7 (0.3) b 12 (1.1) c 10 (0.5) ab 3 (0.3) c 8 (0.9) c 2 (0.4) c 17 (1.3) a 19 (1.2) ab

7 (0.4) b 13 (0.8) c 11 (1.4) ab 5 (0.5) b 9 (0.7) ac 3 (0.4) bc 17 (2.2) a 15 (1.6) c

4 (0.3) ac 3 (0.4) d 7 (0.7) b 2 (0.2) d 3 (0.3) d 1 (0.3) c 7 (0.7) b 11 (0.9) d

5 (0.4) abc 5 (1.5) d 10 (1.3) ab 2 (0.5) cb 4 (0.9) d 2 (0.6) c 9 (2.1) b 16 (1.5) abc

23 93 64 63 100 81 152 140

Total species in 0.1 ha

31 (0.9) ab

38 (2.4) a

33 (2.4) ab

29 (1.6) b

31 (2.4) b

14 (1.1) d

19 (1.8) c

179

Absolute density (number of stems ha 1) Large trees Mid-sized trees Small trees Saplings (ind. m 2) Seedlings (ind. m 2)

244 (27) ac 546 (38) a 792 (146) a 1.3 (0.1) ab 3.9 (0.8) ab

254 (20) a 723 (112) ab 831 (206) a 2.1 (0.4) a 3.4 (1.4) bc

123 (12) b 1027 (267) b 1567 (186) a 1.6 (0.3) ab 1.3 (0.5) b

246 (21) 560 (53) 163 (80) 0.8 (0.1) 5.1 (0.5)

167 (13) 413 (41) 247 (59) 1.6 (0.2) 1.9 (0.4)

67 (10) d 369 (30) c 273 (97) c 0.4 (0.1) c 0.7 (0.1) bc

65 (12) d 310 (47) c 350 (194) acd 0.8 (0.4) bc 1.2 (0.2) d

Total density

53 193 ab

57 733 abc

31 733 ac

60 601 b

35 395 c

13 015 cd

21 420 ab

Basal area for size classes (m2 ha 1) Large trees Mid-sized trees Small trees

46 (2.5) a 16 (1.7) a 4 (0.8) a

37 (4.8) ac 21 (2.9) ab 4 (1.0) ab

16.4 (1.5) bd 31 (7.6) b 9 (0.7) b

40 (3.0) ac 14 (1.4) ac 1 (0.5) c

34 (2.9) c 12 (1.1) cd 1 (0.3) c

10 (1.4) d 11 (1.1) cd 1 (0.5) c

10 (2.5) d 8 (0.9) d 2 (1.3) ac

Basal area for species type (m2 ha 1) Quercus spp. Broad-leaved except Quercus Pinus spp.

17 (2.5) a 50 (2.9) a 0 (0) a

4 (2.3) b 58 (3.7) a 0.5 (0.4) a

0.5 (0.3) c 55 (7.7) a 1.3 (1.1) a

6 (1.4) b 13 (3.9) b 36 (4.1) b

12 (1.7) a 15 (1.9) b 21 (3.0) c

0.7 (0.2) c 2 (0.5) c 19 (1.7) c

0.3 (0.2) c 2 (1.0) c 17 (3.7) bc

Total basal area (m2 ha 1)

67 (2.6) a

62 (3.5) a

56 (9.0) abc

55 (2.7) b

47 (3.1) bc

22 (1.9) d

20 (3.4) d

ac a bcd c ad

b c ad abc bc

a Size classes as in Fig. 1; association names as in legend of Fig. 2. Means within rows followed by different letters are signi®cantly different (Tukey test, p  0:05). No. of sampled plots are given inside curly braces.

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

1a {25}

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

major forest types (Table 3). In general, average number of species (pooled size classes) per plot was signi®cantly higher in QPF (31±38) than in PQLF (29±31), and in PF (14±19; all F6:74 > 3:03, p < 0:01; Table 3). Similarly, total stem density and basal area were higher in QPF than in PQLF, and drastically decreased in the PF stands (Table 3). Mean basal area of Quercus spp., as well as that of other broad-leaved species, was very variable in QPF (0.5±17 m2 ha 1), intermediate in PQLF (6±12 m2 ha 1), and low in PF (0.3±0.7 m2 ha 1; Table 3). Mean basal area of broad-leaved species other than Quercus spp. in QPF was fourfold (50±58 m2 ha 1) that in PQLF (13±15 m2 ha 1), and this latter sixfold that in PF (2 m2 ha 1). Basal area of pine species was higher in PQLF (21±36 m2 ha 1), intermediate in PF (17±19 m2 ha 1), and very low in QPF (0±0.5 m2 ha 1; Table 3). Inverted J-shaped curves of frequency of individuals in diametric size classes were obtained when we pooled species and associations for each of the three

Fig. 3. Number of recorded in the three Quercus±Podocarpus Forest and PF: Pinus

317

stems (ind. ha 1) by diametric categories forest types obtained by TWINSPAN. QPF: Forest; PQLF: Pinus±Quercus±Liquidambar Forest.

major forest types (Fig. 3). The distribution of size classes of QPF was different from PF (Kolmogorov± Smirnov test, two-tailed, p < 0:05), but not from PQLF (p > 0:10); this latter was also not different

Fig. 4. Size distribution for the most important species in the three forest types: (A) QPF; (B) PQLF and (C) PF. Sample size (N), mean of diametric values, and coef®cient of variation (CV) show inequality of diametric size distribution for each species.

318

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

Fig. 4. (Continued ).

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

from PF (p > 0:10). All size classes were represented in QPF and decreased uniformly towards larger classes (>120 cm DBH). Individuals in PQLF were well represented in small and intermediate size classes, but were absent or scarce when DBH > 90 cm. A truncated distribution of size classes was obtained in PF stands, with no individuals with DBH > 82 cm (Fig. 3). The size distributions (>5 cm DBH) of the most important species in each forest type are shown in Fig. 4a±c. Quercus benthamii, Podocarpus matudai and Persea spp. were present in a wide range of stem sizes in the QPF (10±125 cm). The more abundant species in the PQLF were L. styraci¯ua, P. tecunumanii, Q. candicans, and Q. crispipilis, which showed diameters below 85 cm DBH. Eighty percent of total density in PF was accounted by P. oocarpa, whose individuals never reached >82 cm DBH and their mean DBH was relatively low (23 cm). Abundant species in QPF showed a higher size inequality (coef®cient of variation, CV ˆ 0:73 2:16) than those in the PQLF (CV ˆ 0:63 1:44; Mann±Whitney U-test, one-tailed, N1 ˆ N2 ˆ 9, U ˆ 14:5, p ˆ 0:02), and in the PF (CV ˆ 0:71 1:26; Mann±Whitney U-test, one-tailed, N1 ˆ 9, N2 ˆ 4, U ˆ 5, p ˆ 0:04). No differences were found among PQLF and PF (Mann±Whitney U-test, one-tailed, N1 ˆ 9, N2 ˆ 4, U ˆ 17, p ˆ 0:87; Fig. 4a±c).

319

Fig. 5. Frequency of intensity of anthropogenic disturbance in the three forests types. QPF: Quercus±Podocarpus Forest; PQLF: Pinus±Quercus±Liquidambar Forest and PF: Pinus Forest.

3.2. Anthropogenic disturbance The frequency and intensity of anthropogenic disturbance was signi®cantly different among forest types (G-test, w2 ˆ 163:8, d:f: ˆ 6, p < 0:001; Fig. 5). A high proportion of plots (71%) in the QPF did not show evidence of disturbance, as compared to only 24% of the stands in PQLF and 7% in PF. The majority of plots with signs of moderate and severe disturbance were recorded in the PQLF and PF forest types. Most of the disturbance evidence in these plots were related to timber and ®rewood

Fig. 6. Frequency and intensity of anthropogenic disturbance variables recorded in the study sites.

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

320

Table 4 Eigenvalues, accounted variance, and loading for two factors scores and the rated importance of 11 variables included in the factor analysis Extracted factors and their loadings

Percentage of total accounted variance Eigenvalue Input variables Elevation Livestock grazing Timber extraction Resin extraction Fire Basal area of Pinus spp. Soil pH Canopy cover Basal area of Quercus spp. Firewood Slope angle a

Factor 1

Factor 2

48.0 5.27

12.2 1.34

0.966 0.851 0.491 0.705 0.570 0.754 0.778 0.690 0.409 0.631 0.098

0.049 0.042 0.616 0.316 0.542 0.289 0.249 0.156 0.147 0.264 0.016

Communalitya

0.94 0.78 0.78 0.76 0.74 0.72 0.67 0.67 0.51 0.47 0.28

Proportion of variance of variable accounted for by the common factors (see Dillon and Goldstein, 1984).

extraction, and livestock grazing; resin extraction and forest ®res were recorded in 5±15% of the plots (Fig. 6). Factor analysis indicated that anthropogenic disturbance, environmental and structural variables jointly contributed in two common factors to explain more than 60% of total variance (values of communality >0.45, except slope angle; Table 4). The ®rst factor

accounted for 48% of total variation and was positively related with livestock grazing, ®rewood extraction, basal area of Pinus, and high values of soil pH, while being negatively associated with elevation, canopy cover, and basal area of Quercus. The second factor explained an additional 12% of total variation, with timber extraction and forest ®res being inversely correlated.

Fig. 7. Species richness for life forms along ®rst factor ordination. Open circles: QPF, crosses: PQLF, and diamonds: PF.

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

The three major forest types were clearly separated along the axis of the ®rst factor (Fig. 7). A high correlation was found between total number of woody species (r 2 ˆ 0:46, p < 0:001, N ˆ 81) and understory tree species richness (r 2 ˆ 0:72, p < 0:001, N ˆ 81) with the ®rst factor. Species richness of canopy trees and shrubs did not correlate with the ®rst principal factor (r 2 < 0:20, p > 0:10, N ˆ 81; Fig. 7). The highest species richness of understory trees in QPF was positively associated with high elevations and inversely related to frequency and intensity of human disturbance. At the other extreme, PF plots were associated with lower elevation and highest disturbance values. The PQLF plots were distributed in the intermediate range of the ®rst principal factor. 4. Discussion 4.1. Canopy±understory relationships The relationship between forest canopy and understory composition is spatially and temporally dynamic. Forest canopy may determine the composition, structure and functioning of the understory level, which may have reciprocal consequences on regeneration of overstory species through the maintenance of particular conditions in the interior and at the forest ¯oor

321

(Fore et al., 1997). The occurrence of natural disturbances such as landslides, or openings in the canopy, create conditions for recruitment of new individuals or the growth of preexisting ones in the understory (Runkle, 1982, 1990; Pickett and White, 1985; Young and Hubbell, 1991; Attiwill, 1994). However, anthropogenic disturbance, even without removing forest canopy, may also modify recruitment patterns of both canopy and understory species. We found that the number of canopy and understory tree species drastically decreased with increased ranking of human disturbance for the stand. In particular, an impoverished understory tree species layer was associated with a concentration of canopy dominance in a reduced number (2±3) of species of Pinus. Similar ®ndings have been obtained from drier forests in the same study region (GonzaÂlez-Espinosa et al., 1995a,b and unpublished). Pine species are numerous in the greater study region (up to 11 species in the central highlands of Chiapas; Perry, 1991; Farjon and Styles, 1997), but a given stand includes only 2±4 coexisting Pinus spp., with occasional A. xalapensis. Native pines broadly overlap in their successional status, structural attributes and functioning, recruitment and utilization patterns (GonzaÂlez-Espinosa et al., 1997). Therefore, the coexistence of several pine species in the canopy does not necessarily imply that favorable conditions for the maintenance of high species richness in the understory will occur. The

Fig. 8. Geographical distribution of genera (after Mabberley, 1997) and life form in the three forest types. Pantropical does not include genera with neotropical distribution.

322

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

higher number of canopy species in PQLF stands (4±7 species), where pine dominance was highest (plant associations 2a and 2b), was associated to almost twofold decreases of richness of understory trees as compared to QPF. This suggests a negative interaction between pine dominance and diversity of other canopy tree species in explaining local understory richness (Table 3). The three major forest types include differing proportions of ¯oristic elements from different geographical distribution (Fig. 8, based on Mabberley, 1997). The canopy layer of tropical highland forests in southern Mexico and Central America may change its species composition after disturbance towards a larger representation of holarctic genera (e.g. Arbutus, Liquidambar, Pinus, and Quercus). The thinning of the canopy by human activities increases dryness and exposes seedlings and juveniles to more extreme temperatures at the forest ¯oor level. On the other hand, a greater proportion of genera of pantropical or neotropical distribution may be dominant in the understory tree layers of well preserved QPF stands (e.g. Ardisia, Dendropanax, Hedyosmum, Oreopanax, Saurauia, Rapanea, Ternstroemia, and Zinowiewia; Kappelle, 1995; Quintana-Ascencio and GonzaÂlezEspinosa, 1993; Rzedowski, 1993, 1996). A diversity of holarctic elements in the canopy may provide protected understory habitats in which thrive tropical elements in undisturbed forests. However, the understory level of pine-dominated disturbed stands includes a depauperate species richness, mostly including species of holarctic af®nity (e.g. Alnus acuminata, Buddleia spp., Cornus excelsa, Crataegus pubescens, Garrya laurifolia, M. cerifera, Prunus serotina, Rubus spp., Vaccinium spp., and Viburnum jucundum). 4.2. Anthropogenic disturbance and forest structure The primary goal of this study was to determine the role of anthropogenic disturbance in altering ¯oristic composition and forest structure. The interaction between indigenous people and forest ecosystems has existed for centuries in the study region (Collier et al., 1994; Challenger, 1998). The forest has provided the livelihood basis for indigenous social groups which are continuously growing and have been forced to change their resource use strategy, extending themselves into new areas previously occupied by

forests (Morales-HernaÂndez, 2000; Ochoa-Gaona and GonzaÂlez-Espinosa, 2000). The global tendency observed in developing regions is towards satisfying energy needs by means of ®rewood at the cost of over exploitation of the forest (Schulte-Bisping et al., 1999). This trend applies to our study region and does not seem progress towards a positive change in the next future. Firewood extraction and livestock grazing accounted for more of the variance in factor analysis than did timber extraction; however, the role of these sources of disturbance was not similar among forest types. Counting of stumps resulting from forest extraction indicate that the density of small oaks used for ®rewood (<20 cm of DBH) is higher in the PQLF than in the PF, and QPF (338, 238, and 32 ind. ha 1, respectively). The number of relatively large trees harvested for timber (>20 cm DBH), including mostly pine species, varied between 103 ind. ha 1 in the PQLF, 53 ind. ha 1 in the PF, and only 28 ind. ha 1 in the QPF. However, tree extraction in old-growth stands may not only imply a reduction in stock density, but a drastic change in ¯oristic composition as well, as it also includes several highly valued species of oaks and other broad-leaved trees like Podocarpus, Calyptranthes, Clethra, Cleyera, Nyssa, and Prunus. Grazing and trampling in cleared stands causes seedling mortality, in addition to increasing solar radiation and decreasing soil moisture availability at the forest ¯oor level (RamõÂrez-Marcial et al., 1996; Kuiters and Kirby, 1999; Reimoser et al., 1999). These disturbance elements increase the probability of forest ®res during the dry season, which in exceptional years may be as long as 6 months. The frequent occurrence of fog in QPF stands may buffer extreme environmental conditions (Zuill and Lathrop, 1975). The maintenance of many species of oaks is partially favored by its high ®re tolerance (e.g. Parsons and DeBenedetti, 1979; Abrams, 1992; Abrams and Nowacki, 1992; LuÈpke, 1998). However, ®re tolerance is not suf®cient to assure their regeneration after a disturbance. Frequently, even if ®re disturbance is tolerated, it becomes rather unfavorable for the remaining oak trees as it accelerates dead wood extraction by local people. In addition, their removal increases the probability for the establishment of invasive grasses in recently burned forest sites; eventually, grazing is promoted in cleared forest stands.

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

Pine species in the study region are not ®re tolerant and have little resprouting capacity (see Richardson (1998) for cases of pines having reverse responses to ®re). Noss (1999) has suggested that the recovery potential of native forests should not be rated only through using a few species as ecological indicators, but rather with a more holistic approach for the assessment of stand condition. In our study sites, the maintenance of high diversity in old-growth stands cannot be separated from the risks of increasing dominance of the highly successful invading pines. After a period of 96 months since transplant (as 14 months old seedlings) into the ®eld conditions at Huitepec Biological Station in the central highlands of Chiapas, Mexico, Pinus ayacahuite and P. pseudostrobus var. apulcensis survived poorly (4±39%) under closed, undisturbed oldgrowth Quercus canopy. However, their survival increased under a partially cleared canopy (38± 89%) and in open conditions (44±85%; QuintanaAscencio et al., unpublished data). The increase of dominance by native species of pines may delay and prevent the successful colonization of many broadleaved species, whose propagules may not be available or their requirements may not be met in pinelands (Camacho-Cruz et al., in press, report on a ®eld experiment in which a few broad-leaved species could establish when sown on a dense mat of pine litter). We propose that once pines reach a dominance threshold, the possibility of recovery of pinelands into mixed formations may become quite compromised. 5. Conclusion A large portion of the original extent of MRF has been reduced due to agricultural, livestock grazing, and fuelwood and timber extraction. The isolated forest remnants are increasingly deforested as people abandon low productivity agricultural lands under uncertain economic policies. For over 5 decades, Mayan Indian organizations have been recovering territories that were previously under private control, and have established numerous ejidos and new population centers (Burguete, 1999). These social changes have exerted an increasing pressure on forest resources. Uncertainty prevails as to whether the MRF will be able to maintain itself without damage occa-

323

sioned by human activities, and the feasibility of reconverting pinelands, pastures, and abandoned agricultural ®elds into a highly diverse mixed forest (e.g. Hall and Kappelle, 1999; Oosterhoorn and Kappelle, 2000). The effects of land-use patterns on diversity of MRF seem to be highly detrimental, especially for understory trees. Hardwood species are mostly used for fuelwood (even as juveniles), while spontaneously regenerating populations of Pinus spp. are more frequently allowed to attain commercial timber sizes. The enhanced pine dominance in the remaining canopies changes environmental conditions in the interior of the forest and at the forest ¯oor level, and may limit the natural establishment of some species of understory trees (but see Camacho-Cruz et al., in press, for the case of three tree species that could germinate and establish through direct sowing on the forest ¯oor of severely disturbed pinelands). In order to develop ecological restoration practices for pinelands, we need to know more about the biological characteristics and ``silvics'' of at least some key tree species in response to pine dominated canopies. Identifying values for unconventional products (e.g. native hardwoods, phytochemicals, mushroms, ornamental epiphytes; Berlin et al., 1999; Hellier et al., 1999; J.H.D. Wolf, unpublished) in old-growth and mid-successional dense stands, may help to reconcile forest use with regeneration and maintenance of the high diversity of the MRF. Acknowledgements We should like to thank F. BoloÂm-Ton and J.L. LoÂpez-GarcõÂa for their assistance during ®eldwork, and the inhabitants of RincoÂn Chamula and Los Lotes communities for permitting us to access their lands. The owners of La Yerbabuena Reserve provided facilities to accomplish the inventories. We thank R. GonzaÂlez-Montagut, C. MontanÄa, D. Golicher, P.F. Quintana-Ascencio, J. LoÂpez-Portillo, and two anonymous referees for comments that improved the manuscript. The Consejo Nacional de Ciencia y TecnologõÂa (CONACYT) provided a credit-scholarship to NRM (Ref. No. 112676). This research was supported by the European Commission under the INCO-DC programme (Framework 4) as part of the SUCRE

324

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

project (contract No. ERBIC-18CT97-0146), the ComisioÂn Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO, contract No. L-031), and ECOSUR (Federal subsidies 1997±2000). References Abrams, M.D., 1992. Fire and the development of oak forests. BioScience 42, 346±353. Abrams, M.D., Nowacki, G.J., 1992. Historical variation in ®re, oak recruitment, and post-logging accelerated succession in central Pennsylvania. Bull. Torrey Bot. Club 119, 19±28. Aldrich, M., Billington, C., Edwards, M., Laidlaw, R., 1997. Tropical Montane Cloud Forests: an urgent priority for conservation. Biodiv. Bull. No. 2. World Conservation Monitoring Centre, Cambridge, UK. Anonymous, 1998. SPSS Base 8.0. Application Guide. SPSS, Inc., Chicago. Attiwill, P.M., 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. For. Ecol. Manage. 63, 247±300. Bendel, R.B., Higgins, S.S., Teberg, J.E., Pyke, D.A., 1989. Comparison of skewness coef®cient, coef®cient of variation, and Gini coef®cient as inequality measures within populations. Oecologia 78, 394±400. Berlin, B., Berlin, E.A., FernaÂndez Ugalde, J.C., GarcõÂa Barrios, L., Puett, D., Nash, R., GonzaÂlez-Espinosa, M., 1999. The Maya ICBG: drug discovery, medical ethnobiology, and alternative forms of economic development in the highland Maya region of Chiapas, Mexico. Pharm. Biol. 37, 1±19. Breedlove, D.E., 1981. Introduction to the Flora of Chiapas. California Academy of Sciences, San Francisco. Burguete, A. (Ed.), 1999. Mexico: Experiencias de AutonomõÂa IndõÂgena. International King Group for Indigenous Affairs, Copenhague. Burslem, D.F.R.P., Whitmore, T.C., 1999. Species diversity, susceptibility to disturbance and tree population dynamics in tropical rain forest. J. Veg. Sci. 10, 767±776. Camacho-Cruz, A., GonzaÂlez-Espinosa, M., Wolf, J.H.D., de Jong, B.H.J., in press. Germination and survival of trees species in disturbed forests of the highlands of Chiapas, Mexico. Cannon, C.H., Peart, D.R., Leighton, M., Kartawinata, K., 1994. The structure of lowland rain-forest after selective logging in West Kalimantan, Indonesia. For. Ecol. Manage. 67, 49±68. Challenger, A., 1998. UtilizacioÂn y ConservacioÂn de los Ecosistemas Terrestres de MeÂxico. Pasado, Presente y Futuro. CONABIO-Instituto de BiologõÂa-UNAM-AgrupacioÂn Sierra Madre, Distrito Federal, MeÂxico. Churchill, S.P., Balslev, H., Forero, E., Luteyn, J.L. (Eds.), 1995. Biodiversity and Conservation of Neotropical Montane Forests. The New York Botanical Garden, New York. Clark, D.B., 1992. The role of disturbance in the regeneration of neotropical moist forests. In: Bawa, K.S., Hadley, M. (Eds.), Reproductive Ecology of Tropical Forest Plants. Man and the

Biosphere Series, Vol. 7. UNESCO-Parthenon, New York, pp. 291±315. Collier, G.A., 1975. Fields of the Tzotzil: The Ecological Bases of Tradition in Highlands Chiapas. University of Texas Press, Austin, TX. Collier, G.A., 1994. Basta! Land and the Zapatista Rebellion in Chiapas. Institute for Food and Development Policy, Oakland, CA. Collier, G.A., Mountjoy, D.C., Nigh, R.B., 1994. Peasant agriculture and global change. BioScience 44, 398±407. Dillon, W.R., Goldstein, M., 1984. Multivariate Analysis. Methods and Applications. Wiley, New York. Ewel, J., Berish, C., Brown, B., Price, N., Raich, J., 1981. Slashand-burn impacts on a Costa Rican wet forest site. Ecology 62, 816±829. Farjon, A., Styles, B.T., 1997. Pinus (Pinaceae). Flora Neotropica. Monograph 75. The New York Botanical Garden, Bronx. ForeÂ, S.A., Vankat, J.L., Schaefer, R.L., 1997. Temporal variation in the woody understory of an old-growth Fagus±Acer forest and implications for overstorey recruitment. J. Veg. Sci. 8, 607±614. GoÂmez-Pompa, A., VaÂzquez-Yanes, C., 1981. Successional studies of a rain forest in Mexico. In: West, D.C., Shugart, H.H., Botkin, D.B. (Eds.), Forest Succession: Concepts and Application. Springer, Berlin, pp. 246±266. GonzaÂlez-Espinosa, M., Quintana-Ascencio, P.F., RamõÂrez-Marcial, N., GaytaÂn-GuzmaÂn, P., 1991. Secondary succession in disturbed Pinus±Quercus forests of the highlands of Chiapas, MeÂxico. J. Veg. Sci. 2, 351±360. GonzaÂlez-Espinosa, M., Ochoa-Gaona, S., RamõÂrez-Marcial, N., Quintana-Ascencio, P.F., 1995a. Current land-use trends and conservation of old-growth forest habitats in the highlands of Chiapas, Mexico. In: Wilson, M.H, Sader, S.A. (Eds.), Conservation of Neotropical Migrant Birds in Mexico. The Maine Agriculture and Forestry Experiment Station. NSF and USFWS, Orono, pp. 190±198. GonzaÂlez-Espinosa, M., RamõÂrez-Marcial, N., Quintana-Ascencio, P.F., MartõÂnez-IcoÂ, M., 1995b. La utilizacioÂn de los encinos y la conservacioÂn de la biodiversidad en Los Altos de Chiapas. Memorias del III Seminario Nacional sobre utilizacioÂn de encinos. Reporte Cientõ®co UANL, NuÂmero Especial 15. Facultad de Ciencias Forestales, Universidad AutoÂnoma de Nuevo LeoÂn, Linares, N.L., pp. 183±197. GonzaÂlez-Espinosa, M., Ochoa-Gaona, S., RamõÂrez-Marcial, N., Quintana-Ascencio, P.F., 1997. Contexto vegetacional y ¯orõÂstico de la agricultura. In: Parra-VaÂzquez, M.R., DõÂazHernaÂndez, B.M. (Eds.), Los Altos de Chiapas. Agricultura y Crisis Rural. Tomo I. Los Recursos Naturales. El Colegio de la Frontera Sur, San CristoÂbal de Las Casas, Chiapas, MeÂxico, pp. 85±117. GonzaÂlez-HernaÂndez, M., Quintanar, E., 1999. La construccioÂn de la regioÂn autoÂnoma norte y el ejercicio del gobierno municipal. In: Burguete, C.A. (Ed.), Mexico: Experiencias de AutonomõÂa IndõÂgena. IWGIA, Copenhague, pp. 210±233. Hall, K.D., Kappelle, M., 1999. Tropical forest recovery and restoration. Trend Ecol. Evol. 14, 378±379. Hamilton, L.S., Juvik, J.O., Scatena, F.N. (Eds.), 1995. Tropical Montane Cloud Forests. Springer, New York.

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326 Hellier, G.G., Newton, A.C., Ochoa-Gaona, S., 1999. Use of indigenous knowledge for rapidly assessing trends in biodiversity: a case study from Chiapas, Mexico. Biodiv. Conserv. 8, 869±889. Hill, M.O., 1979. TWINSPAN: A FORTRAN program for arranging multivariate data in an ordered two-way table classi®cation of the individuals and attributes. Cornell University, Ithaca, NY. Hong, S.-K., Nakagoshi, N., Kamada, M., 1995. Human impacts on pine-dominated vegetation in rural landscapes in Korea and western Japan. Vegetatio 116, 161±172. Horn, R., Hickey, J., 1991. Ecological sensitivity of Australian rainforests to selective logging. Aust. J. Ecol. 16, 119±129. Hubbell, S.P., Foster, R.B., O'Brien, S.T., Harms, K.E., Condit, R., Wechler, B., Wright, S.J., Loo de Sao, S., 1999. Light-gap disturbances, recruitment limitation, and tree diversity in a neotropical forest. Science 283, 554±557. Kappelle, M., 1995. Ecology of mature and recovering Talamancan montane Quercus forests, Costa Rica. Ph.D. Thesis. University of Amsterdam. Kappelle, M., Geuze, T., Leal, M.E., Cleef, A.M., 1996. Successional age and forest structure in a Costa Rican upper montane Quercus forest. J. Trop. Ecol. 12, 681±698. Kuiters, A.T., Kirby, K.J., 1999. Methodologies in forest grazing research. For. Ecol. Manage. 120, 1±2. LoÂpez-Carmona, M., 1999. Mountainous pastoral system and foraging species in the tzotzil northern region of Chiapas. M.Sc. Thesis. El Colegio de la Frontera Sur, San CristoÂbal de Las Casas, Chiapas. LoÂpez-Portillo, J., Ezcurra, E., 1989. Response of three mangroves to salinity in two geoforms. Funct. Ecol. 3, 355±361. LuÈpke, B. van, 1998. Silvicultural methods of oak regeneration with special respect to shade tolerant mixed species. For. Ecol. Manage. 106, 19±26. Mabberley, D.J., 1997. The Plant Book: A Portable Dictionary of the Vascular Plants. Cambridge University Press, Cambridge, UK. Masaki, T., Tanaka, H., Tanouchi, H., Sakai, T., Nakashizuka, T., 1999. Structure, dynamics and disturbance regime of temperate broad-leaved forests in Japan. J. Veg. Sci. 10, 805±814. Miranda, F., 1952. La VegetacioÂn de Chiapas 1. Ediciones del Gobierno del Estado. Tuxtla GutieÂrrez, Chiapas. Morales-HernaÂndez, M.C., 2000. Uso del suelo y transformacioÂn de los bosques en dos municipios maya-tzotzil de la zona norte de Chiapas, MeÂxico. M.Sc. Thesis. El Colegio de la Frontera Sur, San CristoÂbal de Las Casas, Chiapas. Nagaike, T., Kamitani, T., Nakashisuka, T., 1999. The effect of shelterwood logging on the diversity of plant species in a beech (Fagus crenata) forest in Japan. For. Ecol. Manage. 118, 161±171. Noss, R.F., 1999. Assessing and monitoring forest biodiversity: a suggested framework and indicators. For. Ecol. Manage. 115, 135±146. Ochoa-Gaona, S., GonzaÂlez-Espinosa, M., 2000. Land use and deforestation in the highlands of Chiapas, Mexico. Appl. Geogr. 20, 17±42.

325

Olvera-Vargas, M., Moreno-GoÂmez, S., Figueroa-Rangel, B., 1996. Sitios Permanentes para la InvestigacioÂn SilvõÂcola. Manual para su Establecimiento. Universidad de Guadalajara, Guadalajara, Jalisco. Oosterhoorn, M., Kappelle, M., 2000. Vegetation structure and composition along an interior-edge-exterior gradient in a Costa Rican Montane Cloud Forest. For. Ecol. Manage. 126, 291±307. Parsons, D., DeBenedetti, S., 1979. Impact of ®re suppression on a mixed-conifer forest. For. Ecol. Manage. 2, 21±33. Perry, J.P., 1991. The Pines of Mexico and Central America. Timber Press, Portland. Pickett, S.T.A., White, P.S. (Eds.), 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, New York. Quintana-Ascencio, P.F., GonzaÂlez-Espinosa, M., 1993. A®nidad ®togeogra®ca y papel sucesional de la ¯ora lenÄosa de los bosques de pino-encino de Los Altos de Chiapas. Acta Bot. Mex. 21, 43±57. RamõÂrez-Marcial, N., GonzaÂlez-Espinosa, M., GarcõÂa-Moya, E., 1996. Establecimiento of Pinus spp. y Quercus spp. en matorrales y pastizales de Los Altos de Chiapas, MeÂxico. Agrociencia 30, 249±257. Reimoser, F., Armstrong, H., Suchant, R., 1999. Measuring forest damage of ungulates: what should be considered. For. Ecol. Manage. 120, 47±58. Richardson, D.M. (Ed.), 1998. Ecology and Biogeography of Pinus. Cambridge University Press, Cambridge, UK. Richardson, D.M., Bond, W.J., 1991. Determinants of plant distribution: evidence from pine invasions. Am. Nat. 137, 639±668. Runkle, J.R., 1982. Patterns of disturbance in some old-growth mesic forests of eastern North America. Ecology 63, 1533±1546. Runkle, J.R., 1990. Gap dynamics in an Ohio Acer±Fagus forest and speculations on the geography of disturbance. Can. J. For. Res. 20, 632±641. Rzedowski, J., 1978. VegetacioÂn de MeÂxico. Limusa, Distrito Federal, Mexico. Rzedowski, J., 1993. Diversity and origins of the phanerogamic ¯ora of Mexico. In: Ramamoorty, T.P., Bye, R., Lot, A., Fa, J. (Eds.), Biological Diversity of Mexico: Origins and Distribution. Oxford University Press, New York, pp. 129±144. Rzedowski, J., 1996. AnaÂlisis preliminar de la ¯ora vascular de los bosques meso®los de montanÄa de MeÂxico. Acta Bot. Mex. 35, 25±44. Savage, M., 1997. The role of anthropogenic in¯uences in a mixedconifer forest mortality episode. J. Veg. Sci. 8, 95±104. Scherer, G., Zabowski, D., Java, B., Everett, R., 2000. Timber harvesting residue treatment. Part II. Understory vegetation response. For. Ecol. Manage. 126, 35±50. Schneider, D.W., 1996. Effects of European settlement and land use on regional patterns of similarity among Chesapeake forests. Bull. Torrey Bot. Club 123, 223±239. Schulte-Bisping, H., Bredemeier, M., Beese, F., 1999. Global availability of wood and energy supply from fuelwood and charcoal. Ambio 28, 592±594.

326

N. RamõÂrez-Marcial et al. / Forest Ecology and Management 154 (2001) 311±326

Sheil, D., 1999. Tropical forest diversity, environmental change and species augmentation: the intermediate disturbance hypothesis. J. Veg. Sci. 10, 851±860. Siegel, S., 1956. Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill, New York. Smiet, A.C., 1992. Forest ecology on Java: human impact and vegetation on Montane Forest. J. Trop. Ecol. 8, 129±152. StadtmuÈller, T., 1987. Cloud Forests in the Humid Tropics. A Bibliographic Review. United Nations University, Tokyo and CATIE, Turrialba, Costa Rica. Standley, P.C., Steyermark, J.A., Williams, J.A., Nash, D.L., 1946± 1976. Flora of Guatemala. Fieldiana: Botany 24, Parts I±XII, Chicago. Stapanian, M.A., Cassell, D.L., 1999. Regional frequencies of tree species associated with anthropogenic disturbances in three forest types. For. Ecol. Manage. 117, 241±252.

Veblen, T.T., 1978. Forest preservation in highlands Guatemala. Geogr. Rev. 68, 417±434. Vetaas, O.R., 1997. The effect of canopy disturbance on species richness in a central Himalayan oak forest. Plant Ecol. 132, 29±38. Webster, G.L., 1995. The panorama of neotropical cloud forest. In: Churchill, S.P., Balslev, H., Forero, E., Luteyn, J.L. (Eds.), Biodiversity and Conservation of Neotropical Montane Forests. The New York Botanical Garden, New York, pp. 53±77. Young, T.P., Hubbell, S.P., 1991. Crown asymmetry, treefalls, and repeat disturbance of broad-leaved forest gaps. Ecology 72, 1464±1471. Zuill, H.A., Lathrop, E.W., 1975. The structure and climate of a tropical Montane Rain Forest and an associated temperate pine±oak±Liquidambar forest in the northern highlands of Chiapas, Mexico. Ann. Inst. Biol., Univ. Nac. AutoÂn. MeÂxico, Ser. Bot. 46, pp. 73±118.