A large-scale assessment of ant diversity across the Brazilian Amazon Basin: integrating geographic, ecological, and morphological drivers of sampling bias

Tropical ecosystems are often biodiversity hotspots, and invertebrates
represent the main underrepresented component of diversity in large-scale
analyses. This problem is partly related to the scarcity of data widely
available to conduct these studies and the lack of systematic organization
of knowledge about invertebrates’ distributions in biodiversity hotspots.
Here, we introduce and analyze a comprehensive data compilation of
Amazonian ant diversity. Using records from 1817 to 2020 from both
published and unpublished sources, we describe the diversity and
distribution of ant species in the Brazilian Amazon Basin. Further, using
high-definition images and data from taxonomic publications, we build a
comprehensive database of morphological traits for the ant species that
occur in the region. In total, we recorded 1,067 nominal species in the
Brazilian Amazon Basin, with sampling locations strongly biased by access
routes, urban centers, research institutions, and major infrastructure
projects. Large areas where ant sampling is non-existent represent about
52% of the basin and are concentrated mainly in the North, Southeastern,
and Western Brazilian Amazon. We found that distance to roads is the main
driver of ant sampling in the Amazon. Contrary to our expectations,
morphological traits had lower predictive power in predicting sample bias
than purely geographic variables. However, when geographic predictors were
controlled, habitat stratum and traits contribute to explain the remaining
variance. More species were recorded in better-sampled areas, but species
richness estimation models suggest that areas in South Amazonian edge
forests are associated with especially high species richness. Our results
represent the first trait-based, large-scale study for insects in
Amazonian forests and a starting point for macroecological studies
focusing on insect diversity in the Amazon Basin.

热带生态系统通常属于生物多样性热点区域，而无脊椎动物是大尺度多样性分析中代表性不足的主要类群。该问题的部分成因在于，开展此类研究的公开可用数据较为稀缺，且生物多样性热点区域内的无脊椎动物分布知识缺乏系统性整理。本研究针对亚马逊蚂蚁多样性展开综合数据汇编与分析：我们整合了1817年至2020年间已发表与未发表的观测记录，对巴西亚马逊盆地内的蚂蚁物种多样性与分布格局进行了系统描述。此外，借助高清影像与分类学文献数据，我们为该区域分布的蚂蚁物种构建了涵盖形态性状的综合数据库。最终在巴西亚马逊盆地中共记录到1067个名义物种（nominal species），且采样点位存在显著偏倚，主要集中于通行路线、城市中心、科研机构及大型基建工程周边区域。未开展蚂蚁采样的区域约占该盆地总面积的52%，主要集中在巴西亚马逊北部、东南部与西部区域。研究发现，距道路的距离是影响亚马逊地区蚂蚁采样的核心驱动因素。与研究预期相悖的是，相较于纯地理变量，形态性状对采样偏倚的预测能力较弱。但在控制地理预测变量的影响后，栖息地层位与性状可解释剩余的方差组分。采样强度更高的区域记录到的蚂蚁物种更多，但物种丰富度估算模型显示，亚马逊南部边缘森林区域的物种丰富度尤其可观。本研究成果是首个针对亚马逊森林昆虫的性状驱动型大尺度研究，同时也为聚焦亚马逊盆地昆虫多样性的宏生态学研究提供了重要起点。