Cudney-Valenzuela et al. 2022. Tropical forest loss impoverishes arboreal mammal assemblages through its negative effect on tree canopy connectivity. Figshare dataset.

The Lacandona rainforest is situated in the eastern portion of the State of Chiapas (91˚6’42.8”–90˚41’8.7’’W; 16˚19’17.1”–16˚2’49.3N). Our study was carried out in the Marqués de Comillas region within the Lacandona rainforest, which comprises 203,999 ha of fragmented forest. We selected 20 old-growth forest patches (area ranging from 5 to 2170 ha), separated from each other by at least 2.5 km (distances measured from their geographical centers; Fig. 1).Vegetation structure In the center of each forest patch, we established five 10 × 50 m plots separated by at least 30 m, avoiding forest edges and vegetation gaps. In these plots, we measured the diameter at breast height (DBH) of all trees with DBH ≥ 10 cm, and calculated the sum of tree basal area in each plot. We also estimated the percentage of canopy connectivity inside each plot by taking three hemispherical photographs (25 m apart) with a wide-eye lens (Apexel 198˚ Fisheye Lens). The photos were analyzed using the program Gap Light Analyzer (Frazer et al. 1999) and used the inverse of canopy openness to calculate canopy connectivity. Both tree basal area and canopy connectivity are considered proxies for tree biomass and resource availability. We also considered tree basal area to be a proxy of stand maturity, and canopy connectivity a proxy of canopy movements for arboreal mammals in the focal forest patches.Sampling of arboreal mammals Mammal surveys are detailed elsewhere (Cudney-Valenzuela et al. 2021), but a brief overview is given here. Within the same vegetation plots, we selected five focal trees with suitable climbing conditions (branches ≥ 20 cm wide, preferably hardwood species) and whose architecture allowed installing a camera trap facing other main branches. In each tree, we established a single-rope climbing system. Focal trees in the same patch were separated from each other by a distance ≥ 30 m. Of the five focal trees per patch, four reached the canopy (mean height ± SD = 21.8 ± 6.2 m, range = 10.2 to 36.6 m) and one the midstory (9.1 ± 4.7 m, 3.4 to 19.6 m). Within each patch, we used one camera trap (Bushnell Trophy Cam HD Aggressor Low Glow ©) at a time, which was rotated among the five focal trees once a month, except from October to December when they remained on the same focal tree. Cameras were placed at varying heights depending on the characteristics of the focal tree (camera height of canopy and midstory trees was 15 ± 4.3 m and 2 ± 0.6 m, respectively). Cameras were continuously active from May 2018 to May 2019, and they were serviced once a month (change of batteries, downloading of pictures, replacement of malfunctioning cameras).Total sampling effort was 7,387 camera trap nights (average per patch = 369 ± 11.6 nights), with 6,233 active camera trap nights (average per patch = 311.7 ± 19.9 nights).We calculated each species’ relative abundance index (O’Brien 2011) by dividing the number of events for a given species by the number of days the camera was active in the forest patch, and then multiplying it by 100 . Then we summed the relative abundance index of the species recorded in each patch as an estimate of total abundance of arboreal mammals per forest patch. We used the ‘entropart’ package (Marcon and Hérault 2015) to estimate species diversity using Hill numbers of order 0 (species richness, 0D) and 1 (exponential of Shannon entropy, 1D) (Jost 2006). Species richness (0D) gives disproportionate weight to rare species by considering all species equally abundant while the exponential of Shannon entropy weighs species’ abundances without disproportionately favoring either rare or dominant species, and is therefore interpreted as the number of common (or typical) species in the assemblage (Jost 2006).Forest coverWe adopted a site‐landscape approach (sensu Brennan et al. 2002), in which response variables were measured in same-sized sample sites (i.e., 5 focal trees at the center of each forest patch, total tree basal area and canopy connectivity), and forest cover (i.e., area covered by old-growth forest divided by landscape size × 100) were measured within 13 circular concentric radii (100- to 1300-m radius, at 100 m intervals) from the geographical center of each forest patch (Fig. 1). We used recent and high-resolution Sentinel S2 satellite images (obtained in 2016) to produce land-cover maps of each landscape using ENVI 5.0 software, and extracted forest cover metrics using ArcGIS software with the ‘Patch Analyst’ extension.

拉坎顿雨林（Lacandona rainforest）坐落于墨西哥恰帕斯州东部（西经91°6′42.8″–90°41′8.7″；北纬16°19′17.1″–16°2′49.3″）。本研究在拉坎顿雨林内的马克斯·德·科米亚斯（Marqués de Comillas）区域开展，该区域总面积达203999公顷，以破碎化森林为主。我们共选取20块原生林斑块，斑块面积介于5~2170公顷之间，斑块间地理中心间距不小于2.5km（间隔距离以斑块地理中心测算；图1）。

植被结构：在每块原生林斑块的中心区域，我们设置5个10×50m的样方，样方间距不小于30m，且避开林缘与植被空隙。在上述样方中，我们测定了所有胸径（DBH, diameter at breast height）≥10cm的树木的胸径，并计算了每块样方内的树木胸高断面积总和。此外，我们使用Apexel 198°鱼眼镜头（Apexel 198˚ Fisheye Lens）拍摄3张半球影像（样点间距25m），以此估算每块样方内的冠层连通率。影像通过冠层光隙分析仪（Gap Light Analyzer，Frazer等，1999）进行解析，并以冠层开度的倒数计算冠层连通率。树木胸高断面积与冠层连通率均可作为树木生物量与资源可获得性的替代指标；其中，树木胸高断面积还可作为林分成熟度的替代指标，而冠层连通率则可作为目标斑块内树栖哺乳动物冠层移动通道的替代指标。

树栖哺乳动物采样：哺乳动物调查的详细方法已在既往研究中发表（Cudney-Valenzuela等，2021），本文仅作简要概述。在同一植被样方内，我们选取5棵具备适宜攀爬条件的目标树：枝条直径≥20cm，优先选择硬木树种，且树冠结构便于安装相机陷阱（camera trap）以对准其他主要分枝。每棵目标树上均安装一套单绳攀爬系统。同一斑块内的目标树间距不小于30m。每块斑块的5棵目标树中，4棵可抵达冠层（平均高度±标准差=21.8±6.2m，范围10.2~36.6m），剩余1棵位于中层林（平均高度±标准差=9.1±4.7m，范围3.4~19.6m）。在每块斑块中，我们单次使用1台Bushnell Trophy Cam HD Aggressor Low Glow ©红外相机，每月将相机轮换放置于5棵目标树上，仅10~12月期间相机固定放置于同一目标树。相机安装高度需根据目标树的特性调整：冠层树上的相机安装高度为15±4.3m，中层林树上的相机安装高度为2±0.6m。相机于2018年5月至2019年5月持续运行，每月进行1次维护（更换电池、下载影像、更换故障相机）。总采样工作量为7387个相机陷阱夜（每斑块平均369±11.6夜），有效相机陷阱夜为6233个（每斑块平均311.7±19.9夜）。我们参照O’Brien（2011）的方法计算各物种的相对丰富度指数：将某一物种的探测次数除以该斑块内相机的有效运行天数，再乘以100。随后将每块斑块内记录的所有物种的相对丰富度指数求和，以此作为该森林斑块内树栖哺乳动物总丰富度的估算值。我们使用'entropart'软件包（entropart，Marcon与Hérault，2015）基于希尔数（Hill numbers）估算物种多样性，具体包括阶数0（物种丰富度，0D）与阶数1（香农熵的指数形式，1D）（Jost，2006）。物种丰富度（0D）通过将所有物种视为同等丰富进行计算，对稀有种赋予不成比例的权重；而香农熵的指数形式则依据物种丰度进行加权，不会过度偏向稀有种或优势种，因此可被解释为群落中常见（或典型）物种的数量（Jost，2006）。

森林覆盖度：本研究采用站点-景观方法（Brennan等，2002），其中响应变量在相同尺度的采样位点中测定（即每块原生林斑块中心的5棵目标树、总树木胸高断面积与冠层连通率）；而森林覆盖度（即原生林面积占景观总面积的比例×100）则以每块原生林斑块的地理中心为原点，在13个同心圆形缓冲区（半径范围100~1300m，间隔100m）内测算（图1）。我们使用2016年获取的高分辨率Sentinel S2卫星影像，通过ENVI 5.0软件生成各景观的土地覆盖图，并使用ArcGIS软件搭配'Patch Analyst'扩展模块（Patch Analyst）提取森林覆盖度相关指标。