; ; INTERSPECIFIC SEGREGATION IN A TROPICAL RAIN FOREST AT BAWANGL ING NATURE RESERVE, HAINAN ISLAND

Transcription

1 2003, 27 (3) Ξ Acta Phytoecologica Sinica 3 (, ) hm 2 DBH 1 cm, 2 N N,,, ( < 15 cm) ; ; ( 65 cm) ; 2 ; 2,,, N N INTERSPECIFIC SEGREGATION IN A TROPICAL RAIN FOREST AT BAWANGL ING NATURE RESERVE, HAINAN ISLAND DAI Xiao- Hua YU Shi- Xiao 3 and LIAN Ju- Yu ( School of Life Sciences, Zhongshan University, Guangzhou , China) Abstract Both interspecific association and interspecific segregation can be used to study the spatial affinity between different species. They are connected but different. Interspecific association is more related to habitat, while interspecific segregation more to small- scale effects and intra-/ inter- specific interactions. Association is measured by plot sampling while segregation is measured by plotless nearest neighbor methods. Compared with interspecific association, few studies on pairwise segregation in a multi- species community have been reported. Meanwhile, clarifying the segregation relationship between different species pairs will be helpful for revealing the phenomena of the species interactions, community structure and community dynamics. With the aid of GIS software (ArcView), a distribution map of all trees with DBH 1cm from a tropical rain forest community at Bawangling National Nature Reserve in Hainan Island was drawn. The nearest neigh2 bors of each individual and the distances between every individual- neighbour pair were obtained by using Ar2 cview s nearest-neighbor extension module (Jeff Jenness s Nearest Features v. 3. 5). Based on this, the spa2 tial pattern and interspecific segregation in the multi- species community was studied with a subtable method of a N N nearest-neighbor contingency table. Distribution pattern was measured by a revised Clark- Evans index ( CE) and the often-used 2 method. Pielou s coefficient of segregation ( S) was adopted to measure the seg2 regated degree between two species : while 0. 7 S 1, two species are positively segregated ; - 1 S , negatively segregated ; otherwise they are random neighbors. In order to solve zero cell problems, each zero cell of the contingency table had added to it. This will not change the polarity of segregation and the ze2 ro denominators will be avoided in Pielou s equation. Pairwise segregation was shown in a constellation dia2 gram in which solid lines stand for positive segregation, dash lines for negative segregation and different shapes for different mean DBH. The subtable method is quick, convenient and reliable. Its reliability increases with the increase of number of individuals. This is the reason we only focus on the segregation relationship between those species with more than 10 individuals. With the development of GIS software, computer simulation as well as the application of plotless sampling methods in a large area becomes possible. However, there is not any significant test for Pielou s coefficient until today. The statistical test for such subtables is still under question for both ecologists and statisticians. Ξ : : : ( ) 3 Author for correspondence edu. cn

2 3 : 381 In the forest community we analyzed, some species pairs are positively segregated (23. 3 %) and a few pairs are negatively segregated (2. 4 %). In contrast, the segregation relations between most of the species are random (74. 3 %). Our results support Pielou s opinion that negative segregation is rare in a mature plant community. On the other hand, most understory trees (mean DBH < 15 cm) are positively segregated between each other, few negatively segregated. No negative segregation occurs between large arbors (mean DBH 65 cm). Segregation relations between understory trees and large arbors are complicated. Meanwhile, most species will not be negatively segregated with others. Some will not be positively segregated with others. Two species have both positive and negative segregated relationship, while only one species has neither positive nor negative segregation with all other species. The results also indicated that interspecific segregation is closely re2 lated to their distribution patterns. Proportion of positive segregation between clumped species and other species (31. 4 %) is much larger than that of random species (16. 7 %) or uniform species (20. 0 %). Proportion of negative segregation between clumped species and other species (1. 3 %) is less than that of random (3. 3 %) or uniform species (2. 4 %). Clumped- clumped species pairs tend to be more positively segregated (43. 1 %) than that of other species pairs. Both clumped- uniform and random- random species pairs are more likely to be negatively segregated, with a proportion of 5. 6 % and 4. 8 % respectively. Whatever the species distribution or the pairwise relative distribution is, the proportion of random neighbor pairs is the largest. Possible causes and ecological meanings of interspecific segregation were also discussed in this paper. In2 terspecific segregation results from the combination of many biotic or abiotic factors. For example, negative segregation may occur between dependent unions or those species with similar habitat requirements. Intensive intraspecific competition may result in random or even uniform distribution, therefore such species are apt to be negatively segregated with other species. Large arbors tend to be positively segregated with other large arbors since their size prevents them from occupying a small space. Clumped species should have more conspecific neighbors, that is, less heterospecific neighbors. Thus they have some positive but no negative segregation re2 lationships. Key words Interspecific segregation, Interspecific association, N N nearest-neighbor contingency table, Distribution pattern, GIS ( Interspecific segregation), (Pielou, 1961) ( ( GIS) A B ) ( A (, 1998 ; 2001a) B ) ( ) (Dixon,, 1994) Pielou (1961),, ( Interspecific association) GIS,,, N N ( Pielou, 1961 ; Whipple, 1980 ; Diggle,1983) - ( Hamill & Wright,1986) - (Bawa & Opler,1977) 1 (Pielou, 1961) (Diggle,1983) K ( t) (Ripley,1981) (Dixon,1994),, N, E,, K( t) m, (Coomes et al.,1999) mm, Monte Carlo Reich Davis(1998) N (,1993 ; 2001b ;, N, 2001), 1. 1

3 ; P CE = 1,, 50 m 100 m, CE > 1, CE < 1,, m, ( Kint et al.,2000) :, ( Bischofia javani2 u = r A - r E (2) ca) ( Ficus wightiana) ( Cleistocalyx op2 re erculatus) ( Quercus blakei) ( Firmi2 ana hainanensis) ( Phoebe hungmaoensis) re = = ( Schefflera octophylla) ( Xanthophyllum haina2 N N 2 / A nense) ( Sarcosperma laurinum), r A, r E, N, A (1) ; = N/ A ( Helicia hainanensis ) ; re Poisson r E ( Wrightia hainanensis) ( Walsura ro2 busta) ( Litsea baviensis ), ( Polyalthia laui) ( P. consanguinea) 2 (,1996) ( Ervatamia divaricata ) ( Symplocos d : poilanei) ( Psychortia d = v - 1 (3) rubra) ( Lasianthus hainanensis), d, v 2 ( L. cyanocarpos) d > , ; d < ,, m ;, (Wratten & Fry, 10 m ( 1980) DBH 1 cm) ( ) 1. 3 ( x, y) 1 de Jong (1983) ( Contact sam2 1 m 1 m pling) N N (Subtable GIS (ArcView) DBH 1 method),,reich cm, Davis(1998) N N,, 2,, Pielou (1961) 2 2 ArcView GIS Jeff Jenness s Nearest Fea2, tures v. 3. 5, N N 1. 2 ( 1), Pielou 2 2 Clark Evans(1954), N N i j F ldner (1995) : 2 2 ( 2) 1 CE = r A N N r 2 2 de i i = 1 = Jong (1983) r E 0. 5 A/ N P/ N P/ N 3/ 2 (Short-cut method), 2 (1) Reich Davis (1998), (Subtable) CE Clark- Evans ; r A, 2 ; r E, Pielou (1961) ( Coefficient r A ; r i i of segregation) : ; N ; A S = 1 - N ij ( n ij + n ji ) ( n ii + n ij ) ( n ij + n jj ) + ( n ji + n jj ) ( n ii + n ji ) = 2 ( n ii n jj - n ij n ji ) ( n ii + n ij ) ( n ij + n jj ) + ( n ji + n jj ) ( n ii + n ji ) (4)

4 3 : 383 n ij 0,, n ij :, ;, (3) ;,, (,1996) S , n ii = n jj = 0 n ij,, ; = n ji 0,, S, - 1, ; n ij =, n ji = 0 n ii = n jj 0, ( ),, S - 1, ; n ii n jj = n ij n ji, S = 0, 2 Pielou S (Reich & Davis, 1998) G, 1, : ArcView S 41, 21,18,, 0. 7 S 1,, 2 2 :24-1 S ,17, < S < N N Table 1 N N nearest-neighbor contingency table Base plant Nearest neighbor 1 S 1 2 S 2 3 S 3 k S k Total 1 S 1 n 11 n 12 n 13 n 1k f 1 2 S 2 n 21 n 22 n 23 n 2k f 2 3 S 3 n 31 n 32 n 33 n 3k f 3 k S k n k1 n k2 n k3 n kk f k Total s 1 s 2 s 3 s k N S i : i Species i k : The number of species in the plot n ij : i j Individuals of species i whose nearest neighbor is an individual of species j N : Number of all the individuals f i : i Individuals number of species i s i : i The individuals number of species i as nearest neighbor plants Base plant Table nearest- neighbor contingency table Nearest neighbor Species S i Species S j Total Species S i n ii n ij n ii + n ij Species S j n ji n jj n ji + n jj Total n ii + n ji n ij + n jj N ij N ij : i j, 1 Total number of both species i and species j, the other symbols same as Table 1 ( 65 cm) 1,, ( Garcinia oblongifolia), ( Syzygium brachyantherum), ( Ardisia quinquegona) , 41,820 ; (74. 3 %), ( Ardisia crena2 (23. 3 %), (2. 4 %) ( 4) 1 ta), ;, ( < 15 cm)

5 No Table 3 Distribution patterns of 41 species in a tropical rain forest at Bawangling, Hainan Species Nearest- neighbor method CE u u value 2 Pattern 2 value 2 Chi- square test d d value 1 Psychotria rubra C C 2 Polyalthia consanguinea C C 3 Helicia hainanensis R R 4 Ervatamia divaricata R R 5 Wrightia hainanensis C C 6 Polyalthia laui C R 7 Lasianthus hainanensis C C 8 Symplocos poilanei R C 9 Walsura robusta U R 10 Lasianthus cyanocarpos R R 11 Ardisia crenata R R 12 Litsea baviensis R R 13 Xanthophyllum hainanense R R 14 Aphania oligophylla R R 15 Camellia sinensis var. assamica C C 16 Microcos paniculata R C 17 Dysoxylum binectariferum R R 18 Cleistocalyx operculatus R R 19 Blastus cochinchinensis C C 20 Garcinia oblongifolia C R 21 Schefflera octophylla C C 22 Syzygium brachyantherum R R 23 Memecylon ligustrifolium C C 24 Prismatomeris tetranda C C 25 Sarcosperma laurinum R R 26 Aquilaria sinensis R R 27 Machilus monticola C C 28 Phoebe hungmaoensis R R 29 Oreocnide rubescens C C 30 Machilus sp R R 31 Meliosma angustifolia C R 32 Acer fabri R R 33 Reevesia longipetiolata R R 34 Diospyros cathayensis C R 35 Turpinia montana C C 36 Ardisia obtuse R C 37 Ardisia quinquegona R R 38 Syzygium chunianum U R 39 Ficus harlandii C C 40 Aglaia tsangii R C 41 Litchi chinensis var. euspontanea C R 3 : p < R : Random U : Uniform C : Clumped Pattern 4 Table 4 Proportion for various kinds of segregation between the species with a specific distribution pattern and other species ( %) Distribution pattern Positive segregation Random neighbor Negative segregation Clumped species Random species Uniform species Total

6 3 : Fig. 1 Constellation diagrams of interspecific segregation of 41 species with more than 10 individuals in a ravine rain forest stand at Bawangling 3 Legends to species refer to Table 3 2 Fig. 2 Typical patterns of interspecific segregation a : Negative segregated pair S = (A : Helicia hainanensis B : Xanthophyllum hainanense) b : Random neighbor pair S = (A : Psychotria rubra B : Polyalthia laui) c : Positive segregated pair S = (A : Ardisia crenata B : Polyalthia consanguinea)

7 (2. 4 %) 5,, Clark- Evans (43. 1 %) 4, ; (31. 4 %) (16. 7 %) 5. 6 % (20. 0 %) ;, 4. 8 %, (1. 3 %) (3. 3 %), 5 Table 5 Proportion for various kinds of segregation between the species pairs with different types of distribution pattern ( %) Distribution pattern of species pair Positive segregation Random neighbor Negative segregation 2 Clumped- Clump pair Clumped- Random pair Clumped-Uniform pair Random- Random pair Random-Uniform pair Uniform-Uniform pair Total ),,,, (Pielou,1969) (Law et al.,, 1993) (, ), ( Pielou,1969) Coomes (1999) ( Intermin2 gle) ( Greig-Smith, 1983) Pielou(1961),,, (Non- segregation) ; Coomes, ; ;, 2, ( Pielou,1969),, ( Pielou,, ; 2, 2,,,,,, ;,

8 3 : 387 Greig- Smith, P Quantitative plant ecology. 3rd ed. Oxford : Blackwell Scientific Publications , N Hamill, D. N. & S. J. Wright Testing the dispersion of N, juveniles relative to adults : a new analytic method. Ecology, 67 : ,Kint, V., N. Lust, R. Ferris & A. F. M. Olsthoorn ,, ; Quantification of forest stand structure applied to Scots pine ( Pi2 nus sylvestris L. ) forests. Investigaci n Agraria : Sistemsy Recur2, sos Forestales, 1 : Law, R., A. Mclellan & A. Mahdi Spatio-temporal pro2 cesses in a calcareous grassland. Plant Species Biology, 8 : , 193. Pielou (1961) Lian, J. Y. ( ) & S. X. Yu ( ) Floristic characters of Formation Dacrydium pierrei2syzygium araiocladum : ( Pinus pon2 in the tropical montane rain forest at Bawangling Nature Reserve, derosa) ( Pseudotsuga menzlesii) Hainan Island. Journal of Tropical and Subtropical Botany ( ), 9 : (in Chinese with English ab2, stract) 2 2, Pielou, E. C Segregation and symmetry in two- species pop2 ulations as studied by nearest neighbour relations. Journal of E2, cology, 49 : N Pielou, E. C An introduction to mathematical ecology. New York : John Wiley & Sons N Pielou(1961), Pielou, E. C Mathematical ecology. New York : John Wiley, & Sons Reich, R. M. & R. Davis Quantitative spatial analysis., Fort Collins : Colorado State University , Ripley, B. D Spatial statistics. New York : Wiley. Wang, B. S. ( ), S. X. Yu ( ), S. L. Peng ( GIS ) &M. G. Li ( ) Experimental manual of plant coenology. Guangzhou : Guangdong Higher Education Press. (in Chinese), Whipple, S. A Population dispersion patterns of trees in a, N southern Louisiana hardwood forest. Bulletin of the Torrey Botani2 cal Club, 107 : N, Wratten, S. D. & G. L. A. Fry Field and laboratory exer2 cises in ecology. London : Edward Arnold Yu, S. X. ( ), G. W. Zong( ), Z. Y. Chen( ), R. G. Zang( ) & Y. C. Yang( ) Comparison of ecological entropy with random and systematic Bawa, K. S. & P. A. Opler Spatial relationship between sampling. Acta Phytoecologica Sinica ( ),22 :473 staminate and pistillate plants of dioecious tropic forest trees. Evolution, 31 : Clark, P. J. & F. C. Evans Distance to nearest neighbour as a measure of spatial relationships in populations. Ecology, 35 : Coomes, D. A., M. Rees & L. Turnbull Identifying ag2 gregation and association in fully mapped spatial data. Ecology, 80 : de Jong, P., L. W. Aarssen & R. Turkington The use of contact sampling in studies of association in vegetation. Journal of Ecology, 71 : Diggle, P. J Statistical analysis of spatial point patterns. London : Academic Press Dixon, P Testing spatial segregation using a nearest- neigh2 bor contingency table. Ecology, 75 : (in Chinese with English abstract) Yu, S. X. ( ), H. D. Zhang( ) &B. S. Wang ( ) The tropical montane rain forest of Bawangling Nature Reserve, Hainan Island. I. The ppermanent plots and the community types. Ecological Science ( ),12 (2) : (in Chinese with English abstract) Yu, S. X. ( ), R. G. Zang ( ) & Y. X. Jiang ( ). 2001a. Spatial analysis of species diversity in the tropi2 cal vegetation along the vertical belt at Bawangling Nature Re2 serve, Hainan Island. Acta Ecologica Sinica ( ),21 : (in Chinese with English abstract) Yu, S. X. ( ), R. G. Zang ( ) & Y. X. Jiang ( ). 2001b. Species richness- abundance relationships in four types of tropical forest on altitudinal gradient at Bawangling Na2 ture Reserve, Hainan. Acta Phytoecologica Sinica ( F ldner, K Zur Strukturbeschreibung in Mischbestgnden. ),25 : (in Chinese with English abstract) Forstarchiv, 66 : : :