拉拉山森林動態樣區中林下物種組成與環境的關聯性

Abstract

在亞熱帶森林動態樣區中，針對林下物種組成 (草本、藤本及木本植物) 與環境因子之間關係的研究較少。本研究探討亞熱帶山地雲霧林中環境因子與林下物種組成的關係。為此，我們在拉拉山和塔曼山間的鞍部附近 (北緯 24.2236°、東經 121.4415°) 建立了一個 1 公頃的森林動態樣區，並將其劃分為 100 個 10 m × 10 m 的子樣區。在每個 10 m × 10 m 子樣區的中心，設置 2 m × 2 m 的小樣區，並使用Braun-Blanquet 記錄草本、藤本及胸高直徑小於 1 公分的木本植物的覆蓋度。而環境變量的部分，分為地形 (海拔、凹凸度、坡度、迎風程度及東北向程度)、光照因子 (使用魚眼照片量測)、土壤參數 (土壤 pH 值、土壤深度及土壤含石率) 和小樣區中的地表類型覆蓋度 (土壤、倒木、半倒木及立木)。我們在每一個10 m × 10 m 子樣區裡記錄了樹冠層木本植物 (胸高直徑大於1公分) 物種組成的資料，用以分析它與林下物種組成的關係。對於林下物種組成的分析，使用雙向指示物種分析 (TWINSPAN) 將小樣區分為不同的植群型，判別每種植群型的鑑別種、恆存種與優勢種並比較植群型間的環境差異。我們對地形、光照、土壤參數及棲地覆蓋率這四類環境因子使用了冗餘分析、向前選取法及變異劃分去找出物種組成與環境因子之間的關係。 在小樣區中共記錄到106個物種，包含52種草本植物 (30 種蕨類, 9 種附生植物)、47種木本植物及7種藤本植物。此外，根據 TWINSPAN的結果，林下物種的組成分為兩種不同的植群型：第一種為凸型 (包括66個小樣區)，凸型的小樣區地面具有更高的立木覆蓋度、較低的林下光度；第二種為凹型 (包括34個小樣區)，凹型的小樣區具有較低的立木棲息地覆蓋率、較高的林下光度。另外，凸型的小樣區裡多數的鑑別種為木本植物小苗，而凹型的小樣區裡多數的鑑別種為草本植物。 而環境因子與物種組成的關聯性，根據變異劃分的結果，地形因子中的凹凸度與海拔及棲息地覆蓋度可以解釋較多的總體變量及部分變量。雖然樹冠層物種組成對於林下物種組成有重要的影響，但進一步分析結果發現環境因子與樹冠層物種組成的影響有高度的相關性，因此不能將其從環境因子的影響中獨立出來。對於地形因子的重要性推測為微地形、光照條件、土壤營養成分及微棲地的綜合影響。對於棲息地覆蓋度，分析結果顯示瞭解樣區中微棲地的相對比例的影響對於小尺度的植群研究 (特別是研究林下物種組成) 是很重要的。

Although there are many studies about species composition in subtropical forest dynamics plots, only a few of them focus on the relationship between the species composition in the understory (herbs, lianas, and seedlings of woody species) and environmental variables. In this study, we aimed to explore the relationship between environmental factors and understory vegetation in a subtropical montane cloud forest in Taiwan. We established a 1-ha forest dynamics plot near the saddle between Lalashan and Tamanshan in Northern Taiwan (N 24.2236°, E 121.4415°), divided it into 100 10 m × 10 m subplots, and set 2 m × 2 m quadrat in the center of each subplot. We used Braun-Blanquet scale to record visually estimated coverage of all herbs, lianas, and woody species (with a diameter at breast height less than 1 cm) in each central quadrat. From environmental variables, for 10 m × 10 m subplot we recorded topography (convexity, elevation, northeasterness, slope, and windwardness), light conditions (calculated from photographs taken by fish-eye camera) and soil parameters (rock soil ratio, soil depth and soil pH), and only in each 2 m × 2 m quadrat we also estimated microhabitat type coverage (bare soil type, decayed wood type, half-decayed wood type, and living wood type). In each 10 m × 10 m subplot, we also recorded the overstory woody species composition (DBH ≥ 1 cm), to explore how it influences understory species composition. For understory species composition, we used two-way indicator species analysis (TWINSPAN) to classify the vegetation into types, identified diagnostic, constant, and dominant species of each vegetation type, and quantified environmental differences between them. To find out the relationship between species composition of understory and environmental factors, we used redundancy analysis, forward selection, and variation partitioning among four categories of environmental variables (topography, light, soil variables, and coverage of microhabitat types). To quantify the relationship of understory to overstory, we used ordination axes from detrended correspondence analysis on species composition of overstory as explanatory variables in redundancy analysis on understory species composition. In total, we recorded 106 species, including 52 herbs (30 pteridophytes, 9 epiphyte), 47 woody species, and 7 lianas. According to the TWINSPAN result, the understory species composition was classified into two vegetation types: Type 1 (“Convex”, including 66 quadrats), which has higher coverage of living wood habitat and lower light availability, and Type 2 (“Concave”, including 34 quadrats), which has lower coverage of living wood habitat and higher light availability. Besides, most of the diagnostic species of Convex type are seedling of woody species, whereas, most of diagnostic species of Concave type are herb species. For the species-environment relationship, the result of variation partitioning showed that topographical variables (convexity and elevation) and habitat type coverage can explain more total and partial variance among the four sets of environmental factors. Even though species composition of overstory is thought to be important to understory species composition, according to our results, we couldn’t conclude that overstory vegetation affect the understory vegetation separately because the effect of overstory vegetation was highly correlated with environmental factors. We suggest that the effect of topographical variables combines effects of light conditions, soil nutrients, overstory vegetation, and microhabitat types. Moreover, we also showed that understanding the effect of the relative proportion of microhabitat types in the quadrats is important to fine-scale study, especially studying the understory species composition.