地形对亚热带山地景观尺度植被格局影响的梯度分析

用 7个指标反映三峡大老岭地区森林群落生境的地形特征 ,采用 DCCA排序方法定量分析不同地形因子对亚热带山地森林植被格局的影响及它们各自的生态意义 ;并分析它们的作用强度沿海拔的变化。结果表明 :1)DCCA排序的前 4个轴分别与海拔、坡位和坡度、坡向和坡面显著相关 ;山地森林植被存在多重尺度和方向的分异格局 ;2 )山地森林群落样方生境的热量、水分 (养分 )和光照条件可主要由海拔、坡位和坡度、坡向等地形特征分别反映 ;3)在垂直梯度上 :海拔的影响在海拔 6 0 0～ 80 0 m和 15 0 0～ 170 0 m最大 ;坡向影响的峰值在山顶带 ;坡度、坡位与坡形的影响在山地中部影响最大 ,其中坡度在海拔 6 0 0 m以下的影响也很显著 ;坡面的影响在山地中部最显著。     

祁连山北坡天然草地不同尺度地上生物量空间格局对地形的响应

草地群落地上生物量的空间格局反映了地形差异导致的群落资源配置空间异质性。采用草地群落调查与广义相加模型(GAM)相结合的方法,分析了祁连山北坡典型草原山体尺度和坡面尺度上群落地上生物量格局的响应特征。结果表明:山体尺度上,群落地上生物量受地形因子影响的顺序为海拔坡向变率≈坡度变率≈坡度;利用坡向作为划分条件将山体尺度缩小为坡面尺度后,东、西两个坡面群落地上生物量受坡向变率的影响最大,南、北坡则分别受海拔和坡度变率的影响最大,其中草地群落地上生物量在东坡随坡向变率的增大先减少后增加,西坡随坡向变率的增大先增加后减少,南坡随海拔的上升而增加,北坡随坡度变率的增大先不变后减少;草地群落地上生物量空间格局在山体尺度和坡面尺度上具有明显的尺度依赖特征。     

高寒退化草地狼毒种群地上生物量空间格局对地形的响应

地上生物量的空间格局是物种分布格局的重要内容,在小尺度范围内因地形差异导致的环境异质性是这种格局形成和演变的环境基础.在2011年8月野外调查的基础上,借助ArcGIS和S-PLUS软件,利用广义相加模型(GAM)对退化高寒草地狼毒(Stellera chamaejasme)种群的空间分布格局进行了研究,定量分析了狼毒地上生物量空间格局对主要地形因子的响应机制.结果表明:狼毒种群地上生物量受地形因子影响的顺序为坡向＞坡度＞海拔＞平面曲率＞剖面曲率;坡向和坡度两种地形因子对狼毒的地上生物量空间格局的贡献率分别为3.75和1.48,其余因子则相对较小;狼毒地上生物量在海拔、平面曲率和剖面曲率梯度上的分布比较均衡,在坡向梯度上呈开口向下的抛物线趋势,在坡度梯度上则呈现开口向上的抛物线趋势.狼毒地上生物量空间异质性及其与地形因子之间的关系,反映了狼毒在地形因子对水热条件重分配影响下的响应机制及生长策略.     

基于DEM 的森林景观空间分布分析

在3S 技术的支持下, 通过遥感解译获取南岭国家级自然保护区乳阳片区的森林景观数据, 并建立了该区域数字高程模型(DEM), 提取了海拔、坡度、坡向等地形因子。在ArcGIS9.3 软件中进行叠加分析, 探讨了研究区森林景观空间分布与地形因子(坡度、坡向)的相关关系, 尝试找出其空间分布趋势。结果表明: 阔叶林种对于环境的适应 性是最好的, 在各级坡度和坡向上均有较大范围、较为均衡的分布; 针叶林、山顶阔叶矮林、采伐迹地趋向于聚集分布; 针叶林景观多分布在阳坡和半阳坡, 山地草甸景观倾向于分布在阴坡, 人类干扰所形成的各个景观类型倾向于平坡和缓坡分布。     

庞泉沟国家自然保护区森林群落的数量分类和排序

根据84块样方资料,对庞泉沟国家自然保护区的森林群落进行TWINSPAN分类和DCA、DCCA排序.结果表明：（1）TWINSPAN将该区的森林群落分为7个类型;（2）样方与物种的DCA排序及样方的DCCA排序较好地揭示了该区森林群落的分布格局与环境梯度的关系：DCCA第一轴明显地反映出森林群落的海拔梯度,即热量因子,沿DCCA第一轴从左到右,海拔逐渐升高,森林群落对热量的要求降低;DCCA第二轴主要表现出森林群落的坡向、坡度变化趋势,即水分和光照因子,沿DCCA第二轴,从下到上,坡向由阴坡向阳坡变化,坡度由平缓变陡峭.其中海拔梯度是环境因子中对森林群落分布起决定性作用的因子.（3）与DCA相比,DCCA的排序轴更有利于生态意义的解释,后者能同时反映样方间在种类组成上及环境因子组成上的相似性,表现在排序图中样方较集中,群落间的界线变得较模糊,因此如果同分类结合使用,DCA的效果要好于DCCA.（4）TWINSPAN分类与DCA和DCCA排序的结果,同时表明了该地区森林群落的垂直分布格局.     

中卫山羊核心产地植物群落的数量分类与排序

应用数量分类(TWINSPAN)和DCA排序方法,对我国特有裘皮用山羊(中卫山羊)核心产地——宁夏香山地区植物群落进行多元分析,根据TWINSPAN分类结果,并结合生态特征将所调查的28个样地植被分为6组,代表了该山地在海拔、地形、土壤、山坡坡度等环境作用下植被空间的分布格局。TWINSPAN分类结果与DCA排序结果较一致,DCA排序第一轴结果主要体现出海拔和山坡坡度对植被分布的作用,第二轴结果主要体现了地形(坡向)和土壤基质(沙地、石质山坡、土质山坡)对植被分布的作用。影响该干旱山地植被分布的主要环境因子有海拔和山坡的坡度,另外长期过度放牧对植被的空间分布影响较大,导致了该山地系统植被空间结构紊乱,并干扰了群落数量分析结果的准确性。     

油松人工林林下植物群落变化及其环境解释

通过对山西太岳山宋家沟油松(Pinus tabuliformis)人工林林下植物群落进行调查,采用典范对应分析(CCA),探讨油松人工林林下物种分布与环境因子之间的关系。结果表明:1)在调查的42个样地中,油松林下共出现植物物种83种,隶属26个科72属,其中草本层物种种类较灌木层物种种类丰富,多达64种。2)CCA排序较好地揭示了该区油松人工林林下植物的分布格局与环境梯度的关系:林龄、坡度、坡向、林冠郁闭度、全氮、全钾对油松人工林林下草本层物种分布格局有重要影响;林龄、坡向、坡度、全氮对油松人工林林下灌木层物种分布格局有重要的影响。3)影响林下草本层和灌木层物种分布格局的变量中,生境因子与生物因子共同分别解释了40.3%和42.7%,其中,纯粹的生境变量(即未与生物变量共有的空间变量)对林下植物群落的变异解释能力较生物变量高,表明生境因子在决定油松人工林林下植物分布格局变异中起着举足轻重的作用。     

车八岭山地常绿阔叶林群落结构特征与微地形条件的关系

以车八岭国家级自然保护区南亚热带常绿阔叶林2hm2样地调查数据为基础,根据样方排列、海拔梯度变化及水分梯度理论潜在值从低到高的趋势将样地划分为5条样带,在由山脊至山谷的微地形变化范围内对群落结构进行全面分析,探讨群落内部结构变化规律及其影响因子。研究结果表明,海拔、坡位、坡向、坡度等山地微地形条件对群落结构及物种分布影响显著。从山脊至谷底变化过程中,各样带平均树高、平均胸径以及较大高度级和胸径级的物种个体比例减少,物种多样性呈现增加趋势。多响应置换过程(MRPP)分析表明各样带之间物种组成存在极显著差异(P0.0001),且样带主要物种及密度发生明显变化。通过物种与地形因子的典范对应分析(CCA)知,在同一区域内,物种随地形因子表现出不同的分布趋势,米锥(Castanopsiscarlesii)、甜锥(Castanopsis eyrei)、石栎(Lithocarpus glaber)等喜光的阳性树种在上坡位分布较多,枫香(Liquidambar formosana)、樟叶槭(Acer cinnamomifolium)等喜温暖湿润气候的物种则在下坡位样带为优势物种。揭示了山地常绿阔叶林群落特征对微地形变化的响应,对于区域生物多样性保护和植被恢复具有重要的理论和实践意义。     

神农架米心水青冈-曼青冈群落的地形异质性及其生态影响

地貌单元和地形的局部变化严重制约着光照、温度、水分甚至土壤资源的再分配 ,从而对植被的分布产生影响。以永久性监测样地为基础 ,利用数字高程模型模拟 0 .96 hm2样地的地形表面 ,在地理信息系统的支持下分析了米心水青冈 -曼青冈群落的地形异质性状况 ,并结合群落调查的结果探讨了地形因素对群落组成的影响。结果表明研究范围内样方的平均坡度在 18.2～ 74 .6°之间。坡向可以分为南坡、西南坡及东南坡 3部分 ,坡度及坡向差异显著。对坡度、坡向等地形因素及群落的物种组成进行半方差分析 ,发现三者的空间异质性尺度分别为 2 85 .8m,2 80 .5 m及 6 .1m。地形因素的变化尺度远大于植被组成的变化 ,说明小尺度下地形因素不是影响群落内种群分布格局的唯一因素。坡度和坡向因子的块金值分别为 2 1.8及 4 98.0 ,说明两者在小尺度下的某些过程不能忽视。植物与地形因素的相关分析表明不同的种对地形的适应性不同 ,群落中的 2 4种主要植物可以分为 4个生态种组。因此 ,群落的物种组成变化受到多种因素的影响 ,这种影响可以在地形变化中得到反映。     

黑河上游天然草地蝗虫物种丰富度与地形关系的GAM分析

地形差异性导致的环境异质性为小尺度范围内生物空间格局的形成与维持提供了一种重要机制, 是形成物种丰富度差异性的前提条件。借助GIS和S-PLUS软件, 利用广义可加模型(GAM)于7-8月对影响蝗虫分布的地形因子进行了研究, 在定量分析黑河上游祁连山区北坡地形的海拔分异特征的基础上研究了该区域蝗虫的丰富度与地形复杂度的关系。结果表明： 在36个样方中共采集蝗虫3 149头, 隶属于3科10属13种; 蝗虫丰富度受地形因子影响的顺序为海拔＞坡向＞坡度＞剖面曲率＞平面曲率＞坡位; 蝗虫的分布在平面曲率和剖面曲率各个梯度上的分布比较均衡, 在海拔、 坡向以及坡位的每个梯度上呈二次抛物线分布, 坡度上呈递减趋势; 从分布的区域上来看, 蝗虫在整个区域都有较高的丰富度, 但主要分布在海拔2 600~2 700 m区域, 坡向上则主要集中在西北坡和西坡, 与实际观测情况相一致。蝗虫丰富度与地形因子之间的相互关系以及分布状态, 反映了地形因子对水热条件的重分配使蝗虫分布格局出现多元化以及破碎化。     

人为干扰对风水林群落林下木本植物组成和多样性的影响

华南地区的风水林是乡村聚落的一种特色林分, 具有守护村庄的象征意义。在过去的数百年中, 风水林在乡村的社会文化习俗的影响下而受到保护, 对当地的生物多样性保育有着重要作用。为揭示人为干扰对风水林的影响, 我们选择广东省东莞市大岭山镇同一林分起源的3个具有相似地形的风水林, 研究了在不同干扰强度下其林下木本植物种类组成和物种多样性。多响应置换过程(multi-response permutation procedures, MRPP)分析表明, 人为干扰显著改变了风水林林下木本植物组成(P = 0.001, A = 0.3886), 沿着干扰由弱至强的梯度呈现出中生性植物减少、阳生性植物递增的趋势。多样性指数变化趋势为重度干扰＞中度干扰＞轻度干扰, 但没有表现出统计学意义上的差异(P＞0.05)。随着干扰强度的增大, 3个风水林群落相互间的林下物种相似性降低, 物种替代率呈增加趋势。双向聚类分析较好地反映出林下物种因受不同人为干扰强度影响而表现出在空间分布上的差异。指示种分析进一步确定了不同干扰强度下具有显著指示值(IV ≥60)的指示种。综合分析表明, 人为干扰有利于阳性物种在风水林内定居生长, 并明显地改变了林下木本植物组成, 但未能引起物种多样性的显著差异。此外, 找出对人为干扰产生关键生态响应的林下指示种, 对增进风水林的生物多样性保育以及生态系统管理有着重要的理论意义和实践价值。     

A multivariate model of plant species richness in forested systems: old-growth montane forests with a long history of fire

Recently, efforts to develop multivariate models of plant species richness have been extended to include systems where trees play important roles as overstory elements mediating the influences of environment and disturbance on understory richness. We used structural equation modeling to examine the relationship of understory vascular plant species richness to understory abundance, forest structure, topographic slope, and surface fire history in lower montane forests on the North Rim of Grand Canyon National Park, USA based on data from eighty‐two 0.1 ha plots. The questions of primary interest in this analysis were: (1) to what degree are influences of trees on understory richness mediated by effects on understory abundance? (2) To what degree are influences of fire history on richness mediated by effects on trees and/or understory abundance? (3) Can the influences of fire history on this system be related simply to time‐since‐fire or are there unique influences associated with long‐term fire frequency? The results we obtained are consistent with the following inferences. First, it appears that pine trees had a strong inhibitory effect on the abundance of understory plants, which in turn led to lower understory species richness. Second, richness declined over time since the last fire. This pattern appears to result from several processes, including (1) a post‐fire stimulation of germination, (2) a decline in understory abundance, and (3) an increase over time in pine abundance (which indirectly leads to reduced richness). Finally, once time‐since‐fire was statistically controlled, it was seen that areas with higher fire frequency have lower richness than expected, which appears to result from negative effects on understory abundance, possibly by depletions of soil nutrients from repeated surface fire. Overall, it appears that at large temporal and spatial scales, surface fire plays an important and complex role in structuring understory plant communities in old‐growth montane forests. These results show how multivariate models of herbaceous richness can be expanded to apply to forested systems.     

Woody plant communities of isolated Afromontane cloud forests in Taita Hills,Kenya

In the Taita Hills in southern Kenya, remnants of the original Afromontane forest vegetation are restricted to isolated mountain peaks. To assess the level of degradation and the need for forest restoration, we examined how forest plant communities and their indicator species vary between and within remnant patches of cloud forest. We used ordinal abundance data to compare plant communities in eight forest fragments. We also analyzed data on the diversity and abundance of trees in 57 0.1 ha plots to compare tree communities within and between the largest two of these fragments, Ngangao (120 ha) and Mbololo (220 ha). The extant vegetation of the Taita Hills at landscape scale consists of secondary moist montane to intermediate montane forest. There was a high species dissimilarity between fragments (69%). Variation in species composition coincided with an abiotic gradient related to elevation. At plot level, secondary successional species and species of forest edges were most abundant and most frequent. Inferred clusters of plots almost entirely coincided with the two forest fragments. Indicator species associated with forest margins and gaps were more frequent in the smaller of the two forest fragments, while indicators for the larger fragment were more typical for less disturbed moist forest. Abiotic site variability but also different levels of disturbance determine site-specific variants of the montane forest. Conservation efforts should not only focus on maintaining forest quantity (size), but also on forest quality (species composition). Late-successional rainforest species are underrepresented in the woody plant communities of the Taita Hills and assisting restoration of viable populations of cloud forest climax tree species is urgently needed.     

The Spatial-Temporal Heterogeneity of Understory Light Availability in a Temperate Forest of North China

The spatial-temporal variation of understory light availability has important influences on species diversity and community assembly. However, the distribution characteristics and influencing factors of understory light availability have not been fully elucidated, especially in temperate deciduous, broad-leaved forests. In this study, the understory light availability was monitored monthly (May–October) in a temperate deciduous, broad-leaved forest in Henan Province, China. Differences in the light availability among different months and habitat types were statistically analyzed using Kruskal–Wallis method, respectively. Partial least squares path modeling (PLS-PM) was used to explore the direct and/or indirect effects of stand structure, dominant species and topographic factors on the light environment. Results showed that there were differences in light environments among the four habitat types and during the studied six months. The PLS-PM results showed that the stand structure and the dominant species were negatively correlated with the light environment, and the path coefficient values were −0.089 (P = 0.042) and −0.130 (P = 0.004), respectively. Our result indicated that the understory light availability exhibit a distinct spatial and temporal heterogeneity in temperate deciduous, broad-leaved forest of north China. The characteristics of woody plant community, especially the abundance of one of the dominant plant species, were the important factors affecting the understory light availability.     

Moth assemblages in Costa Rica rain forest mirror small-scale topographic heterogeneity

In many tropical lowland rain forests, topographic variation increases environmental heterogeneity, thus contributing to the extraordinary biodiversity of tropical lowland forests. While a growing number of studies have addressed effects of topographic differences on tropical insect communities at regional scales (e.g., along extensive elevational gradients), surprisingly little is known about topographic effects at smaller spatial scales. The present study investigates moth assemblages in a topographically heterogeneous lowland rain forest landscape, at distances of less than a few hundred meters, in the Golfo Dulce region (SW Costa Rica). Three moth lineages—Erebidae–Arctiinae (tiger and lichen moths), the bombycoid complex, and Geometridae (inchworm moths)—were examined by means of automatic light traps in three different forest types: creek forest, slope forest, and ridge forest. Altogether, 6,543 individuals of 419 species were observed. Moth assemblages differed significantly between the three forest types regarding species richness, total abundance, and species composition. Moth richness and abundance increased more than fourfold and eightfold from creek over slope to ridge forest sites. All three taxonomic units showed identical biodiversity patterns, notwithstanding their strong differences in multiple eco-morphological traits. An indicator species analysis revealed that most species identified as characteristic were associated either with the ridge forest alone or with ridge plus slope forests, but very few with the creek forest. Despite their mobility, local moth assemblages are highly differentially filtered from the same regional species pool. Hence, variation in environmental factors significantly affects assemblages of tropical moth species at small spatial scales.     

广西凭祥不同年龄红椎林林下植物物种多样性及其环境解释

林下植被作为森林生态系统的重要组成部分,其在调控森林群落结构和功能上发挥着至关重要的作用。从揭示人工林林下植物物种组成的关键控制因素的目标出发,以不同年龄红椎林灌木层和草本层的植物群落作为研究对象,采用野外调查和室内分析相结合的方法,重点探究林下植物物种组成与环境因子之间的关系及其关键控制因子。研究结果表明:不同年龄红椎林的林下植物物种组成存在明显差异;RDA结果表明,胸高断面积是影响灌木层物种组成的最主要因子,胸高断面积、坡度和海拔能显著影响影响草本层的物种组成;方差分解结果表明,所调查的生物和非生物因子分别解释灌木层和草本层物种组成变异的63%和47%,微环境和植物之间的相互作用是影响林下植物群落物种变异的最主要因素,而土壤与微环境、土壤与植物两者之间,以及微环境、植物和土壤三者之间的相互作用对林下植物群落的物种组成的影响较小,甚至无影响,微环境的独立效应对林下植物物种组成的影响高于植物或土壤因素。     

三峡大老岭山地常绿落叶阔叶混交林林隙干扰研究II.林隙干扰的地形格局

根据对三峡大老岭地区山地常绿落叶阔叶林林隙的调查结果,运用多种方法定量分析和检验了5种地形要素对林隙和3类林隙形成木(GM)各方面特征的空间格局的影响。结果反映了地形对林隙时空格局的影响,和不同类型林隙形成木的响应特征：1）林隙干扰随海拔上升、坡度增大、坡位从山脊到沟谷、坡形由凸到凹而加剧;坡向影响的规律不明显。林隙特征对各地形因素梯度有显著的响应,其中扩展林隙面积、林隙高度、GM数量、种数以及平均胸径是反映地形影响的敏感指标。2）GM的平均胸径、最大胸径和扩展林隙面积与海拔显著正相关;GM平均胸径还与坡位正相关;GM数量沿N→S的坡向梯度减少,而随坡度增大及坡形由凹转凸而增加;GM种类沿坡度和坡形梯度有相同的变化趋势。林隙高度（亦即植被高度）随海拔和坡度增大而减小,但随坡位上升而增大。地形要素对林隙高度和GM平均胸径的空间格局的贡献最大,分别达到75.9%和67.0%。3）折干特征的空间格局受海拔和坡位的显著影响;枯立木的分布特征主要受小地形的坡位和坡形控制;影响翻倒木的最根本地形因素是坡度。     

西双版纳热带雨林林窗空间分布格局及其特征数与林窗下植物多样性的相关性

林窗作为森林群落中一种重要的干扰方式, 对林下物种构成有着重要的影响。开展林窗空间格局及其特征指数与林下植物多样性关系研究对于探讨林窗对林下生物多样性的影响有重要意义, 有助于进一步了解群落动态, 在物种多样性保护方面也具有指导作用。本研究在西双版纳热带雨林地区随机选取3块大小为1 ha的热带雨林为研究样地, 采用轻小型六旋翼无人机搭载Sony ILCE-A7r可见光传感器, 分别获取各个样地的高清数字影像, 结合数字表面高程模型以及各个样地的地形数据用以确定各样区的林窗分布格局, 并进一步提取出各林窗的景观格局指数。结合地面样方基础调查数据, 对各样地各林窗下植物多样性情况进行统计, 旨在分析热带雨林林窗空间分布格局以及林窗下植物多样性对各林窗空间格局特征的响应情况。研究表明, 西双版纳州热带雨林林窗呈大而分散的空间分布, 林窗空间格局特征指数如林窗形状复杂性指数、林窗面积都与林下植物多样性呈显著正相关关系。在面积小的林窗下, 较之林窗形状复杂性因子, 林窗面积大小对林下植物多样性影响更显著; 在面积达到一定程度后, 相对于面积因子, 林窗形状复杂性指数对林下植物多样性影响更显著, 各样地林窗皆趋于向各自所处样地顶极群落发展。     

山西灵空山小蛇沟林下草本层植物群落梯度分析及环境解释

通过对山西灵空山小蛇沟集水区的林下草本层植物群落进行调查和多元分析——TWINSPAN分类、典范对应分析(CCA)与生境、生物因素变量分离, 探讨林分水平上草本层物种分布与环境因子之间的关系。结果如下: 1) TWINSPAN将26个调查样方划分为6种群落类型: 以辽东栎(Quercus wutaishanica)为主的辽东栎-油松(Pinus tabulaeformis)林型、辽东栎杂木林型、辽东栎林型、华北落叶松(Larix principis-rupprechtii)林型、油松林和阔叶油松林型、油松-辽东栎均匀混交林型, 体现了该地区地带性植被类型为暖温带森林的特点。2)群落类型的划分与CCA的结果相吻合, 主要反映了CCA排序第一、二轴的环境梯度, CCA排序轴第一轴突出反映了林分类型与土壤养分梯度, 第二排序轴与坡度、坡位显著相关。Monte Carlo检验结果表明, 林分类型、土壤养分和坡度是影响小蛇沟集水区内林下草本物种分异的最主要的环境因子。3)生境因子与生物因子解释了物种格局变化的42.9%, 其中生境因子占31.8%, 生物因子占7.9%, 生境因子与生物因子交互作用解释部分占3.2%。良好的环境解释反映了调查取样和环境因子选取的合理性。对于50%以上未能被解释的变异部分, 可能归咎于未被选取的因子如干扰或者随机过程。4)在海拔梯度较小的山区, 坡向等小地形因子能较好地指示局部生境的小气候条件, 对林下植物的分布有较好的解释力。     

Tree Diversity,Forest Structure and Productivity along Altitudinal and Topographical Gradients in a Species-Rich Ecuadorian Montane Rain Forest

We studied the spatial heterogeneity of tree diversity, and of forest structure and productivity in a highly diverse tropical mountain area in southern Ecuador with the aim of understanding the causes of the large variation in these parameters. Two major environmental gradients, elevation and topography, representing a broad range of climatic and edaphic site conditions, were analyzed. We found the highest species richness of trees in valleys <2100 m. Valleys showed highest values of basal area, leaf area index and tree basal area increment as well. Tree diversity also increased from ridges to valleys, while canopy openness decreased. Significant relationships existed between tree diversity and soil parameters (pH, total contents of Mg, K, Ca, N and P), and between diversity and the spatial variability of pH and Ca and Mg contents suggesting a dependence of tree diversity on both absolute levels and on the small-scale heterogeneity of soil nutrient availability. Tree diversity and basal area increment were positively correlated, partly because both are similarly affected by soil conditions. We conclude that the extraordinarily high tree species richness in the area is primarily caused by three factors: (1) the existence of steep altitudinal and topographic gradients in a rather limited area creating a small-scale mosaic of edaphically different habitats; (2) the intermingling of Amazonian lowland plant species, that reach their upper distribution limits, and of montane forest species; and (3) the geographical position of the study area between the humid eastern Andean slope and the dry interandean forests of South Ecuador.