Abstract Habitat loss and fragmentation are among the main threats to global biodiversity and ecosystem services. However, major research biases and knowledge shortfalls in some geographical regions, taxonomic groups and responses studied are recurrent in fragmentation-related research. Therefore, detecting these biases and associated gaps is crucial to steer future research efforts and to guide applicable conservation policies. Here we conducted an exhaustive literature review to evaluate biases on fragmentation research across biomes, taxonomic groups, species responses and fragmentation metrics in Brazil. Overall, we analysed 716 papers, comprising a database with 26 taxonomic groups and 1173 cases studied across the six Brazilian biomes. In general, we observed that fragmentation-related research was biogeographically biased towards forest biomes. Specifically, the Atlantic Forest, the most populated and deforested Brazilian biome, comprised the highest number of studies (63%), while non-forest biomes were largely underrepresented. We also detected a high positive relative taxonomic bias for birds and mammals, while many insect taxa were disproportionately underrepresented in the literature. Altogether, assemblage-level species responses (abundance, diversity and richness) comprised 72% of study cases. Moreover, fragment size was clearly the most considered metric in the studies (43%) followed by habitat quality and edge effects. Our findings indicate major information deficits with regard to fragmentation-related research among taxonomic groups and amongst biomes in a megadiverse country. Therefore, we suggest that fragmentation research conducted in Brazil needs to consider undersampled taxa and to be urgently extended to increasingly degraded non-forest biomes in order to avoid inappropriate inferences.

中文翻译：

---

巴西碎片化研究的主要偏见和知识差距：对保护的影响

摘要 栖息地丧失和破碎化是全球生物多样性和生态系统服务面临的主要威胁之一。然而，一些地理区域、分类群和所研究的反应的主要研究偏见和知识短缺在与碎片相关的研究中反复出现。因此，检测这些偏见和相关差距对于指导未来的研究工作和指导适用的保护政策至关重要。在这里，我们进行了详尽的文献综述，以评估巴西跨生物群落、分类群、物种响应和碎片化指标的碎片化研究的偏见。总体而言，我们分析了 716 篇论文，包括一个数据库，其中包含 26 个分类群和在六个巴西生物群落中研究的 1173 个案例。一般来说，我们观察到与碎片相关的研究在生物地理学上偏向于森林生物群落。具体而言，大西洋森林是巴西人口最多和森林砍伐最多的生物群落，其研究数量最多 (63%)，而非森林生物群落的代表性不足。我们还检测到鸟类和哺乳动物的高正相对分类偏差，而许多昆虫分类群在文献中的代表性不足。总而言之，组合水平的物种反应（丰富度、多样性和丰富度）占研究案例的 72%。此外，碎片大小显然是研究中考虑最多的指标（43%），其次是栖息地质量和边缘效应。我们的研究结果表明，在一个多样化国家的分类群之间和生物群落之间，与碎片相关的研究存在重大信息不足。