Understanding which species will be extirpated in the aftermath of large‐scale human disturbance is critical to mitigating biodiversity loss, particularly in hyperdiverse tropical biomes. Deforestation is the strongest driver of contemporary local extinctions in tropical forests but may occur at different tempos. The 2 most extensive tropical forest biomes in South America—the Atlantic Forest and the Amazon—have experienced historically divergent pathways of habitat loss and biodiversity decay, providing a unique case study to investigate rates of local species persistence on a single continent. We quantified medium‐ to large‐bodied mammal species persistence across these biomes to elucidate how landscape configuration affects their persistence and associated ecological functions. We collected occurrence data for 617 assemblages of medium‐ to large‐bodied mammal species (>1 kg) in the Atlantic Forest and the Amazon. Analyzing natural habitat cover based on satellite data (1985–2022), we employed descriptive statistics and generalized linear models (GLMs) to investigate ecospecies occurrence patterns in relation to habitat cover across the landscapes. The subregional erosion of Amazonian mammal assemblage diversity since the 1970s mirrors that observed since the colonial conquest of the Atlantic Forest, given that 52.8% of all Amazonian mammals are now on a similar trajectory. Four out of 5 large mammals in the Atlantic Forest were prone to extirpation, whereas 53% of Amazonian mammals were vulnerable to extirpation. Greater natural habitat cover increased the persistence likelihood of ecospecies in both biomes. These trends reflected a median local species loss 63.9% higher in the Atlantic Forest than in the Amazon, which appears to be moving toward a turning point of forest habitat loss and degradation. The contrasting trajectories of species persistence in the Amazon and Atlantic Forest domains underscore the importance of considering historical habitat loss pathways and regional biodiversity erosion in conservation strategies. By focusing on landscape configuration and identifying essential ecological functions associated with large vertebrate species, conservation planning and management practices can be better informed.

中文翻译：

---

利用历史栖息地丧失来预测新热带森林中当代哺乳动物的灭绝

了解哪些物种将在大规模人类干扰后被灭绝对于减轻生物多样性丧失至关重要，特别是在高度多样化的热带生物群落中。森林砍伐是热带森林当代局部灭绝的最有力驱动因素，但可能以不同的速度发生。南美洲两个最广泛的热带森林生物群落——大西洋森林和亚马逊——在历史上经历了不同的栖息地丧失和生物多样性衰退途径，为调查当地物种在单一大陆上的持续存在率提供了独特的案例研究。我们量化了这些生物群落中的中型到大型哺乳动物物种的持久性，以阐明景观配置如何影响它们的持久性和相关的生态功能。我们收集了大西洋森林和亚马逊地区 617 个中型至大型哺乳动物物种（> 1 公斤）的出现数据。我们根据卫星数据（1985-2022）分析自然栖息地覆盖，采用描述性统计和广义线性模型（GLM）来调查与整个景观的栖息地覆盖相关的生态物种发生模式。自 20 世纪 70 年代以来，亚马逊地区哺乳动物群落多样性的次区域侵蚀反映了自殖民征服大西洋森林以来所观察到的情况，因为现在 52.8% 的亚马逊哺乳动物都处于类似的轨迹。大西洋森林中五分之四的大型哺乳动物容易灭绝，而亚马逊地区的哺乳动物有 53% 容易灭绝。更大的自然栖息地覆盖增加了两个生物群落中生态物种持续存在的可能性。这些趋势反映了大西洋森林当地物种损失的中位数比亚马逊高出 63.9%，这似乎正在走向森林栖息地丧失和退化的转折点。亚马逊和大西洋森林地区物种持久性的对比轨迹强调了在保护战略中考虑历史栖息地丧失路径和区域生物多样性侵蚀的重要性。通过关注景观配置并确定与大型脊椎动物物种相关的基本生态功能，可以更好地了解保护规划和管理实践。