Abstract: [Objective]To further explore the species coexistence mechanisms in tropical mountain rainforest, we compared differences in species composition and population characteristics in different topographic conditions, and assessed topographic associations with species diversity.[Method]The 60 hm² tropical montane rainforest in Jianfengling of Hainan Province containing trees (DBH ≥ 1 cm) of 290 woody species as a large plot was classified into different topographic types according to their elevations, convexities and slopes of each of the 20 m×20 m sample plots by fuzzy C-mean clustering. community composition, species abundance, and individual density among different topographic conditions were compared. Using Torus-translation tests, we examined the association between micro-topography and species for 230 woody species each with more than 60 plants with a DBH ≥ 1 cm in the plot.[Results]All sample plots of Jianfengling plot could be unambiguously assigned to one of the four topography categories, lower valley (12.6 hm², 274 species), steep slope (16.68 hm², 269 species), upper valley (15.08 hm², 264 species) and ridge (15.64 hm², 267 species). By comparing species-area relationships among these four topographies, the number of species in the lower valley was the largest. But there was little difference among other three topographies, because the numbers of species of them fell within the expected values of the steep slope ±1.96×SE. By comparing the cumulative species-individual relationships among these four topographies, the cumulative rate of the lower valley was the largest and it was the least of the ridge. But there was little difference between the steep slope and the lower valley, because the numbers of species of them fell within the significance interval of the expected values of steep slope ±1.96×SE. The upper valley had the largest average DBH, followed by the steep slope, whereas the lower valley had the least average DBH. And the average DBH of the ridge was equal to that of the whole plot. The ridge had the highest density, followed by the upper valley. The density of the tree community of the upper valley was lower than that of the lower valley, when the average DBH ≤ 10 cm. When the average DBH ≥ 10 cm, it was inverse. For the proportion of different DBH sizes, the upper valley > the steep slope > the ridge > the lower valley when the average DBH ≥ 30 cm, and the lower valley > the upper valley > the steep slope > the ridge when the average DBH ≥ 50 cm. Of the lower valley type, Gironniera subaequalis, Blastus cochinchinensis, and Livistona saribus were the most dominant species, while of the other three topographies Livistona saribus, Gironniera subaequalis, and Crypcarya chinensis were the most dominant species. 203 (88.3% of all) of 230 common species showed significant associations with the topographies, which is relatively higher than other large plots in the world. We used the mean of species negatively associated with one of the four topographies divided by the examined 230 species to assess the effects of topography. The extent that species-topography associations contributed to species coexistence was 23.7%.[Conclusion]We found that community structure, species diversity and species-topography associations were different among 4 topographies in the 60 hm² plot in Jianfengling. Topography was one of the important environmental factors that determined species distribution in the plot, although light, moisture and soil condition may also have played roles. This study indicated that the species in Jianfengling plot had strong differentiation of micro-topographic niche. Topographic heterogeneity had important contribution to maintaining species diversity in this tropical montane rainforest. This study also provided guidance to choose different species combinations in different topographies when restoring tropical mountain rainforest, rebuilding or constructing mixed forests, according to the species-topography associations.

Key words: tropical montane rainforest, topography, species distribution, Torus-translation test, habitat associations of woody plants, topographic heterogeneity, species coexistence