BIOTROPIA Vol. 28 No. 1,2021: 29 - 37 DOI: 10.11598/btb.2021.28.1.907 DOMINANCE, ASSOCIATION AND DISTRIBUTION PATTERN OF TREE SPECIES I N BURNT FOREST I N EAST KALIMANTAN SUBEKTI RAHAYU', AGUS PRIYONO IOZRTON02, SAMBAS BASUNI2 AND AGUS HII(MAT2 ' WorldAgraforesbty, Jalan C@r, Situ Gede, Sindang Barang, Bogor 16680, Indonesia 'Forest Resources Conseruntion and Ecotourism Department, Faculg $Foresty, Institut Pertanian Bogor, Kavzpzrs IPB Dramaga, Bogor 16680, Indonesia Received 4 October 2017/Accepted 17 February 2020 ABSTRACT Repeated forest fues remarkably impact species composition. Pioneer species colonize the burnt forest and widely develop up to 30 years after a fire but late-succession species regenerate gradually or even disappear owing to direct impact of fires or other ecological consequences related to fires. Hence, forest restoration through assisted natural regeneration needs some information about the state of post-fire species composition. T o better evaluate tree species composition after repeated fires, the species dominance as an indicator of species composition was used in this research, with additional information o n the species association and distribution patterns. A 1.8-hectare plot, divided into 180 subplots with size of 10 x 10 m, was established in a secondary forest in Samboja Research Forest, East Icalimantan. The sample plot was burnt in 1982/1983 and 1997/1998. All trees above 10 cm DBH were measured and leaf specimens were collected for species identification at the Herbarium Bogoriense, Cibinong, West Java. For comparison, the 1981 data from the Indonesian Institute of Science were used. Analysis of the Importance Value Index used the species dominance data. A 2 x 2 matrix based o n the presence-absence of species for each subplot was used to analyse the association index among species. Variance and average value ratio of certain species present in each subplot were expressed in a dispersion index. A chi-square was used to test the significance between the association and dispersion index. Thirteen years after a second fire, pioneer species of Macarangagigantea were most dominant, followed by Vernonia arborea, a sub- climax species. This indicated that the forest was in an early succession process. Pholidocarpw mqadtm was consistently dominant before and after the fire. A total of 38 pairs of species were significantly positively associated and 4801 pairs negatively associated. About 60% of species association, both negatively and positively, were among the 'native species' (species that existed before the fire events) and 'non-native species' (new comer species that regenerated after the fire) in the plot sample. A non-native species, Vernonia arborea, associated negatively with the non-native species Tabernaemontana sphaerocarpa, and native species, Oncospem homdum, Palaquizlm das_yph_vllzm and Endiandra mbescens. The distribution pattern of four native species, Artocarpzls anisoph_vlhs, Cananga odorata, Croton laev~olius and Macarangagigantea, changed after repeated fires, from uniform to clumped. Keywords: association, distribution pattern, dominance, East Icalimantan, forest fire INTRODUCTION Samboja Research Forest is a remnant, lowland, mixed, Dipterocarp forest ecosystem in East Icalimantan, Indonesia, in whch the commercial and high-quality timber species grow, such as meranti (Shorea sp.) and ulin (Ezlsideroydon ~agem) (Delmy 2001). Fifty-five (55) species of Dipterocarpaceae were preiously *Corresponding author, email: s.rahayu@cgiar.org growing in the area (I E(a), that is, the pair of species occurred together more often than expected if independent. 2. Negative, if a < E(a), that is, the pair of species occurred together less often than expected if independent. Index of dispersion OD) used to express the species distribution pattern based on the variation to mean ratio (Ludwig & Reynolds 1988), was computed as: where: o2 is the variance and p is the mean The three types of dispersion index are namely; (1) random, if o2 = p, (2) clumped, if o2 > p and (3) uniform, if o2 < p. Chi-square statistics used to test significantly different ID, was computed as: Degree of freedom (db) is N-1. RESULTS AND DISCUSSION Species Dominance The five most dominant species observed in the sample plots of the 10.5 hectare undisturbed ulin and Dipterocarp forest in Samboja Research Forest in 1981 were Sborea laevis, Pbolidocarpw mdjadun, Dio@yros borneensis, Ezlsideroylon zpvagem' and Scapium macropopdurn (Icartawinata et al. 2008). During the 1981 observation, three most dominant species were found in the smaller plot of 1.8 ha of undisturbed Pbolidocarpas majadam) Dio@yros borneensis and Eztsideroylon pagem'. Further analysis was focused on the smaller sample plots in the 1.8 ha. Five years after the second fire of 1997/1998, three most dominant species based on stem basal area were found in the smaller plots, namely; Pbolidocarpas myadan) Easideroylon pagem' and Dipterocarpas cornatas (Simbolon 2005). These three species had individual trees that survived the repeated fires. The surviving individual trees were recognised from tree mapping analysis through comparing coordinate position and stem diameter of 1981 and 2011 data. Large surviving trees having > 40 cm diameter of Easideroylon pvagem' and Dipterocarpas cornatas) as well as the palm species, Pbolidocarpas mdjadan, having 20 - 30 cm diameter contributed both high basal area and IVI. Other species that started to establish during the period had contributed to small IVI. The burnt area was dominated by trees having < 5 cm diameter and 10 - 15 cm tree diameter (Sirnbolon 2005). Thirteen years after the second fire (2011 observation), the three most dominant species dramatically changed from the climax species of Ezasideroylon pagem' and Dipterocarpas cornata~ to pioneer species of Macaranga gigdntea and sub- climax species of Vernonia arborea. Pbolidocarpus mdjadan was consistently in the third place Fable 1). The ten most dominant species were mostly pioneer, except Ezlsideroylon ?wagem' and Dipterocarpw cornatas. This clearly indicated that 13 years after a fire disturbance, pioneer species colonized the open area. However, surviving species continued to grow among the pioneer species. The surviving individuals of Pbolidocarpas mdjadan provided h g h levels of seed for the regeneration process. This endemic palm in Icalimantan, mostly growing in swamp forest in the study area, was unaffected by the fires (Simbolon 2005). Moreover, mortality of this palm was relatively low, reaching 10% in the burnt forest, probably due to the physiological characteristics of its stem vascular structure (van Nieuwstadt & Shiel2005). Dominance, association and distribution pattern of tree species in burn forest in East Icalimantan - Rahayu e t al. Table 1 Species with the 10 highest M at 1 3 years after the second fire (2011 observation) No. Family Species Local name Importance Value Index 1 Euphorbiaceae Macaranga gigantea Merkubung 35.29 2 Asteraceae Vernonia arborea Merambung 30.54 3 Arecaceae Pholidocarpus majadum Liran 12.98 4 Moraceae Artocarpus anisoph_ylhs Mentawa 12.27 5 Rutaceae Melicope glabra Sampang/terutup 12.19 6 Euphorbiaceae Croton laevfoli,vs Lasa-lasa/belanti 7.46 7 Lauraceae Eusideroylon .ywagen' Ulin 7.20 8 Verbenaceae Peronema canescens Sungkai 6.71 9 Annonaceae Cananga odorata ICenanga 5.73 10 Dipterocarpaceae Dipterocarpw cornutus Keruing gajah 5.62 Easideroqlon pagem' is another surviving species that had a slight decrease in M from 9.3 before to 7.2 after the fire. E. ?wagem' has the capacity to re-sprout from damaged trees after fire (Delmy 2001), even from stumps and roots (van Nieuwstadt & Shiel 2005). This species is categorized as heavy wood, ranging 0.88 - 1.19 g/cm3, on average 1.04 g/cm3 Wartawijaya e t al. 1992). Heavy wood species (> 0.8 g/cm3) have the capacity to survive fire (van Nieuwstadt & Sheil 2005; Brando e t a l 2012). High density wood produces some extractive material, such as cellulose, hemicelluloses, and lignin, and other properties, that reduce the fire distribution index and ignition time (Brando e t al. 2012). Before fire, Macaranga gigdntea was a minor species in the permanent plot of Samboja Research Forest (Icartawinata e t a/. 2008). After repeated fires, the number of trees increased rapidly from 35 in 2003 (Simbolon 2005) to 167 in 2011. Three years after the fire, Macaranga gigantea was the most dominant species in the twice-burned forest, (Slik e t aL. 2008) and was consistently dominant at 5, 7 and 13 years. This species was regularly found as one of the most dominant species after 10 - 20 years in regenerating forest (Silk e t aL. 2008). However, in Samboja Research Forest, M . ggigdntea was the most dominant species three years after the fire. The level of disturbance after the fire might have affected the early establishment of Macaranga gigdntea. Macaranga is an early succession genus that prefers a frequently disturbed habitat that is characteristic of a highly-disturbed forest (Slik e t al. 2003). M. gigdntea traits of being a pioneer and light demanding species are suitable for growth in burnt forest where high exposure to sunlight occurs. Its typically orthodox seeds are able to lie dormant in the soil and germinate as soon as the forest canopy opens up (Suita & Nurhasybi 2009; Susanto e t a l 201 6). Vemonia arborea, a sub-climax species (Desitarani e t al. 2015), was not found in the sample plot before the fire events but developed rapidly and became the second dominant species in 201 1. However, t h s species was found in an undisturbed forest located at 1 - 2 krn from the sample plot (JGisnawati e t aL. 201 1, Atmoko e t a l 2015). The small seeds of V: arborea can be dispersed by wind. Moreover, this species was a co-dominant in the once-burned forest of Sungai Wein but not in the twice-burned Samboja (Slik e t al. 2008b). Increasing A13+ after fires probably affected V. arborea colonization in burnt areas. A study in Samboja Lestari, East Icalimantan documented that Al+ content in the soil significantly increased four years after a fire and I/. arborea was associated with high A13+ content (J'asir e t al. 2010). High A13+ content in the soil becomes a limiting factor of root growth (Mossor-Pietraszewska 2001) for common species, but not for V. arborea whch is adapted to high A13+. In Samboja, the V. arborea population developed widely after 2003 (five years after the second fire). The occurrence of pioneer and s u b - c h a x species as the most dominant, together with establishing climax species in the burned forest has indicated that the succession process was in a competition phase (Clement 191 6). In order to address restoration management goals, the assisted natural regeneration and species enrichment activities can be applied in the forest succession phase (Elhot e t al. 2013; Lamb & Gilmour 2006). Assisted natural regeneration should focus on species priorities and enrichment, particularly for sub-climax and climax species, to improve the forest function and therefore, conserve biodiversity. BIOTROPIA Vol. 28 No. 1,2021 Species Association negative association occurred more frequently Thirteen years after the second fire, an analysis of the 191 established species in the sample plots of secondary forest showed that 18,336 pairs of species pairing with themselves, consisted of 1,220 (7%) pairs with positive association, 17,115 (93%) pairs with negative association and a pair of species as neutral. Chi- square test found 38 pairs with significantly positive association and 4,801 pairs with negative association. T h s positive association rarely occurred but negative association was common, particularly, among 'non-native' species (Icuebbing & Nunez 2015). Non-native and native species in this research refers to the existence of species before and after the fires. Non-native species were found in the plot only after the fires but native species were found before and after the fires. However, positive association is important in establishing management priorities (Icuebbing & Nunez 2015). I t is critical in community structure or habitat modification, in which one individual or species alters the condition of the local environment, often making a stressful habitat more hospitable to other individuals or species (Stachowicz 2001). Furthermore, positive association between species in mutual interaction will enhance each other's survival probabilities (Ludwig & Reynolds 1988). Positive interactions between different species are of particular interest because of their potential to 'cascade' throughout the com- munity, with a major effect on the ecosystem structure and function (Stachowicz 2001). Some 53% positive association and 60% negative association in the sample plot occurred between species that existed before the fire ('native') and newcomer species that regenerated after the fire ('non-native'). In this research, the between 'native7 and 'non-native' rathe; than among 'non-native'. The more frequent occurring association among the 'non-native' species indicated that 'non-native' species indirectly provide environmental support (Flory & Bauer 201 4). Improving performance of the 'non-native' species after fires is associated with natural resources availability or changes in disturbance regime (Daehler 2003). However, light competition as an indirect impact of fire disturbance was an important factor affecting interaction among individual trees or species (Icunstler e t al. 2012). Vernonia arborea, a 'non-native' species, was positively associated with one 'native7 species Anthocephalm chinensis but negatively associated with other 'native7 species, Oncospem homa'um, Palaquium d a q p b l h m , E didndm mbescens and 'non-native7 species, Tabernaemontana sphaerocapa (Table 2). A. chinensis, a surviving individual tree provided a suitable environment for sub-climax species of V. arborea seedlings. Anthocephal,w chinensis grows in the swampy area of the plot unaffected by fires (Simbolon 2005). V. arborea generally grew in the upper elevations of the sample plot. Only a few individual trees grew on the swamp area close to Anthocephalzls chinensis. As a sub-climax species, V. arborea probably took advantage of A. chinensis' canopy for shade during early regeneration. Negative association between I/. arborea with Tabernaemontana sphaerocapa, Oncospem hom'dum, Palaquium daqpbllzm and Endiandra mbescens was probably because of different habitat requirements. Tabernaemontana sphaerocarpa, Oncospem bom'dum, Palaquium daqpkyllum and Endidndm mbescens grew on lower elevations, in the swampy part of the sample plot. Table 2 Ten species' pairs with highest Yates Chi-square value Pair of species Yates x2 Critical Association A Ea df correction value type Knema lateriGia Neonauclea cabcina 2 0.07 15 32.24 25.00 + Calopbylhm soulattri Di0.pyro.r pilosanthera 2 0.08 8 25.59 15.51 + Anthocephalus chinensis Vernonia arborea 4 2.92 12 24.31 21.03 + Knema cinerea Pobalthia mmphii 2 0.09 3 23.80 7.82 + Pobalthia m p h i i Q p g i m brachyrachis 2 0.09 3 23.80 7.82 + Tabernaemontana sphaerocarpa Vernonia arborea 0 0.83 2 1861.84 5.99 Tabernaemontana hauilandii Vitex pinnata 0 0.01 2 129.56 3.84 Oncosperma homdum Vernonia arborea 0 0.83 2 127.65 3.84 Palaquium dasyphyllzm Vernonia arborea 2 2.50 2 91.47 25.00 Endiandra mbescens Vernonia arborea 0 0.83 2 68.78 3.84 Notes: A= number of subplots where both species, A and B, are present; Ea= Expected value of a species associated with other species. Dominance, association and distribution pattern of tree species in burn forest in East Icalimantan - Rahayu et al. V. arborea generally grew in the upper elevations of the sample plot. Only a few individual trees grew on the swamp area close to AnthocephaZas chinensis. As a sub-climax species, V. arborea probably took advantage of A . chinensis' canopy for shade during early regeneration. Negative association between V. arborea with Tabernaemontana sphaerocarpa, Oncosperma hom'dam, PaIaqaiam dagpbIIam and Endiandm rabescens was probably because of chfferent habitat requirements. Tabernaemontana .phaerocarpa, Oncospema homzmzd~m, PaZaqaiam dagtp&lZam and Endidndm mbescens grew on lower elevations, in the swampy part of the sample plot. Species Distribution Pattern species with significantly distributed clumped (Table 3). The changing distribution pattern of Artocarpzls anisopbIZus, Cananga odo~ata, Croton Zaevfolizls and Macaranga gigdntea from uniform to clumped, after repeated fires, was due to various factors, such as tree survival in the cooler site and their stem size (Davis et al. 2005). The survivor trees in the cooler site functioned as seed sources when most trees were killed, while their small stem size tended to clump. The nature of its fruits and seeds had also affected their distribution pattern after the fire. Artocarpas anisop&Zlzs has a compound fruit. It contains many 1.7 x 1.0 cm seeds that are difficult to chsperse without a dispersal agent, such as a big mammal. The swampy area was suitable for - Cananga odorata and Croton Zaevfohas. Swampy In the 1981 observation, the species distribution in the 1.8-hectare sample plot of area depletion and droughts after the fires unchsturbed forest was a balance between became a limiting factor for C. odorata and C. uniform (48%), where the distance between Iaevfohas growing in the sample plot. Macaranga neighbouring inchviduals is maximised due to gigdntea, a minor species and distributed competition of resources, and random (52%) uniforml~, changed dramatically to dominant where the spacing between individual is and clumped. After biomass burning, open areas unpredictable, usually occurring in the habitat were associated with pioneer species, such as M. with consistent environmental condition. gigdfftea, due to hgh light as an Repeated fires affected the species' distribution pattern. Analysis of 191 species in the 1.8 hectare secondary growth forest indicated that 69 species (369'0) were distributed randomly, 28 species (14'/0) were clumped and 95 species (50%) were uniformly distributed. Clumped dispersion of species was found in the secondary growth forest after repeated fires. Further statistical testing using the Chi-square found six source during the colonization process. Clumped pattern of Peronema canescens might affect the planting of this species in a forest rehabilitation program. During the forest rehabilitation activities in Samboja Researh Forest in 1990s P. canescens was one of the species planted in the area for fire breaks and boundary area signs to community land. Then, it has spread out after the repeated forest fire. Table 3 Dispersion index of six species in the 1.8-hectare secondary forest after repeated fires in Samboja Research Forest Species name Dispersion Variance Average Index d f Critical value Adocarpus anisap~llz.r 1.97 0.37 5.28 348.74 66 85.98 Cananga odorata 0.21 0.1 3 1.66 36.61 22 32.67 Croton laevfolizls 0.31 0.1 8 1.77 54.86 31 44.98 Macaranga gigantea 1.37 0.93 1.48 245.95 166 197.06 Peronema canescens 0.33 0.1 8 1.80 57.48 32 46.19 Vernonia arborea 1.29 0.75 1.73 231.32 134 162.01 BIOTROPIA Vol. 28 No. 1,2021 CONCLUSION The dominance of Macaranga gigantea indicated a high level of forest disturbance in that part of Samboja Research Forest affected by two fire events. Co-dominance of Vemonia arborea indicated the changing environmental conditions, such as increasing A13+, whch was limiting to the species growth. Occurrence of pioneer, sub-climax and climax species indicated that the forest succession was in a competition phase. Assisted natural regeneration and species enrichment would be possible restoration activities for biodiversity conservation. Improving non-native species performance in the twice-burned area might occur due to changing natural resources availabihty and &sturbance regime after a fire. Lastly, the changing environmental conditions, species characteristics and human activities affected the distribution pattern of species. REFERENCES Atrnoko T, Yassir I, Sitepu BS, Mukhlisi, Widuri SA, Muslim T, Mediawati I, Ma'mf A. 2015. Keanekaragaman Hayati Hutan Rintis Wartono Icadri: Hutan Tropis I G m a n t a n di IWDTIC Samboja. Balikpapan (ID): Balai Penelitian Teknologi Konservasi Sumber Daya Alam. 104p. Brando PM, Nepstad DC, Balch JI<, Bolker B, Christman MC, Coe M, Putz FE. 2012. Fire-induced tree mortality in a neotropical forest: the roles of bark traits, tree size, wood density and fire behavior. Global Change Biology 18:630-41. Delmy A. 2001. Fire resistant of tree species in Bukit Soeharto Education Forest, East Kalimantan, Indonesia. In: Icobayasi e t al., Editors. Rehabilitation of Tropical Forest Ecosystems. Bogor (ID): Center for International Forestry Research. Desitarani, Wiriadinana H, Miyakawa H, Rachman I, Rugayah, Sulistyono, Partomihardjo T. 2015. Buku panduan lapangan jenis-jenis tumbuhan restorasi. Jakarta (ID): Kementerian Lingkungan Hidup dan Icehutanan-JICA-Lembaga Ilmu Pengetahuan Indonesia. Dobson AP, Bradshaw AD, Baker AJM. 1997. Hopes for the future: restoration ecology and conservation biology. Sciences 277:515-20. Elliot SD, Blakesly D , Hardwick K. 2013. Restoring tropical forests: a practical guide. Richmond (UIC): Royal Botanic Gardens Ice, 344p. Flory SL, Bauer J. 2014. Experimental evidence for indirect facilitation among invasive plants. J Ecol 102:12-8. Janzen D H . 1970. Herbivores and the number of tree species in tropical forests. American Naturalist 104(940):501-28. Kartawinata I<. 2010. Dua abad memgungkap kekayaan flora dan ekosistem Indonesia. Presented in Sanvono Prawirohardjo Memorial Lecture X, Indonesia Institute of Science, Jakarta. 23 August 2010. Bogor (ID): LIPI. IGrtawinata I<, Purwaningsih, Partomihardjo T, Yusuf R, Abdulhadi R, Riswan S. 2008. Floristic and a structure of low land Dipterocarp forest at Wanariset Samboja, East Kalimantan, Indonesia. Reinwartia 12(4) :301-23. IGisnawati H, Wahjono D, Imanuddin R. 2011. Changes in the species composition, stand structures and aboveground biomass of a lowland Dipterocarps forest of Samboja, East Kalimantan. J For Res Clement FE. 1916. Plant succession: an analysis of the 8(1):1-16. Washingt0n DC (US): Kuebbing SE, Nunez MA. 2015. Negative, neutral, and Carnegie Institution of Washington. positive interactions among nonnative plants: Condit R, Pitman N , Leigh E G Jr, Chave J, Terborgh J , patterns, processes, and management implications. Foster RB. 2002. Beta-diversity in tropical forest Glob Chang Biol21:926-34. trees. Science 295:666-9. Kunstler G, Lavergne S, Courbaud B, Thuiller W, Curtis JT. 1959. The Vegetation of Wisconsin: An Ordination of Plant Communities. Madison (US): University of Wisconsin Press. Davis MA, Curran C, Tietmeyer A, Miller A. 2005. Dynamic tree aggregation pattern in a species-poor temperate woodland disturbed by fire. J Veg Sci 16:167-74. Daehler CC. 2003. Performance comparison of co- occuring native and alien invasive plants: implication for conservation and restoration. Annu Rev Ecol Evol Systc 3 4 1 83-211. Vieilledent ~ y ~ i r n m e r m a n n NE, . . ., Coomes DA. 2012. Competitive interactions between forest trees are driven by species' trait hierarchy, not phylogenetic or functional similarity: implications for forest community assembly. Ecol Lett 15: 831-40. Lamb D , Gilmour D . 2003. Rehabilitation and restoration of degraded forests. Gland (CH): IUCN. Li L, Ye WH, Wei SG, Lian JY, Huang ZL. 2014. Spatial patterns and associations between species belonging to four genera of the Lauraceae Family. PLoS One 9(11): e l 11 500. Dominance, association and distribution pattern of tree species in burn forest in East Icalimantan - Rahayu e t al. Ludwig JA, J F Reynolds. 1988. Statistical ecology: A primer on methods and computing. New York (US): John Wiley & Sons, Inc. Martawijaya A, Kartasujana I, I