Pervasive Defaunation of Forest Remnants in a Tropical Biodiversity Hotspot

Tropical deforestation and forest fragmentation are among the most important biodiversity conservation issues worldwide, yet local extinctions of millions of animal and plant populations stranded in unprotected forest remnants remain poorly explained. Here, we report unprecedented rates of local extinctions of medium to large-bodied mammals in one of the world's most important tropical biodiversity hotspots. We scrutinized 8,846 person-years of local knowledge to derive patch occupancy data for 18 mammal species within 196 forest patches across a 252,669-km2 study region of the Brazilian Atlantic Forest. We uncovered a staggering rate of local extinctions in the mammal fauna, with only 767 from a possible 3,528 populations still persisting. On average, forest patches retained 3.9 out of 18 potential species occupancies, and geographic ranges had contracted to 0–14.4% of their former distributions, including five large-bodied species that had been extirpated at a regional scale. Forest fragments were highly accessible to hunters and exposed to edge effects and fires, thereby severely diminishing the predictive power of species-area relationships, with the power model explaining only ∼9% of the variation in species richness per patch. Hence, conventional species-area curves provided over-optimistic estimates of species persistence in that most forest fragments had lost species at a much faster rate than predicted by habitat loss alone.

热带森林砍伐与森林破碎化是全球生物多样性保护的核心议题之一，但滞留在未保护森林残余斑块中的数百万动植物种群的局地灭绝现象，至今仍未得到充分阐释。本研究聚焦全球最重要的热带生物多样性热点（biodiversity hotspots）区域之一，报道了中大型哺乳类前所未有的局地灭绝速率。我们梳理了总计8846人年的本土调研资料，在面积达252669平方千米的巴西大西洋森林（Brazilian Atlantic Forest）研究区域内，获取了196个森林斑块中18种哺乳动物的斑块占用数据。研究团队发现该区域哺乳类群的局地灭绝速率触目惊心：原本预估存在的3528个种群中，仅767个仍存续。平均而言，每个森林斑块仅保留了18个潜在物种中的3.9个；物种的地理分布范围已缩减至原本分布的0%~14.4%，其中5个大型物种已在区域尺度上完全灭绝。森林斑块极易被猎人进入，且面临边缘效应与火灾的干扰，这极大削弱了种-面积关系（species-area relationships）的预测能力：幂模型（power model）仅能解释约9%的单个斑块物种丰富度变异。因此，常规种-面积曲线对物种存续的预估过于乐观——绝大多数森林斑块的物种消失速率远快于仅由栖息地丧失所预测的水平。