Data from: Biotic interactions in species distribution models enhance model performance and shed light on natural history of rare birds: a case study using the Straight-billed Reedhaunter (Limnoctites rectirostris)

Species distribution models (SDMs) have become a workhorse to explain, understand and predict distributions of birds. However, SDMs at broad scales are typically built using climatic variables, while ignoring the effects of biotic interactions. Although its role still remains controversial, the inclusion of biotic interactions into SDMs could confirm and/or provide new ecological insights of poorly known species. We modeled the distribution of the rare South American straight-billed reedhaunter (Limnoctites rectirostris, Furnariidae), a specialist of marshy areas linked to the spiny herb eryngo (Eryngium spp., Apiaceae), which provides the main food and nest resources. To do this, we first modeled the distribution of three eryngo species considered as the main biotic interactors (E. eburneum, E. horridum and E. pandanifolium) and included them into the straight-billed reedhaunter SDM. Second, we analyzed niche overlap between the straight-billed reedhaunter and eryngos in terms of environmental variables using dynamic range boxes, a novel approach to quantify size of n-dimensional hypervolumes. The inclusion of biotic interactions improved model performance relative to a model with climatic variables only. Climatic suitability of E. eburneum and mean temperature of wettest quarter were the most important predictors. By contrast, E horridum and E. pandanifolium resulted in poor predictors, suggesting that the straight-billed reedhaunter’s relative dependence on each eryngo species is different. The three eryngo environmental spaces largely covered the environmental space of the straight-billed reedhaunter, but the opposite was not true. Our findings suggest that biotic interactions play an important role in explaining and predicting the distribution of a rare bird at macro-scales, and that the assessment of niche overlap between interactors may confirm or improve the autoecological understanding of rare and cryptic birds. We advocate the use of an integrative modeling approach including climate and biotic interactions into SDMs to enhance ecological knowledge of poorly known bird species.

物种分布模型（Species Distribution Models，SDMs）现已成为解释、认知与预测鸟类分布的核心工具。然而，大尺度下的物种分布模型通常仅基于气候变量构建，忽略了生物交互作用的影响。尽管生物交互作用的生态功能仍存在争议，但将其纳入物种分布模型，可为认知匮乏的物种提供新的生态学视角，或验证已有认知。
本研究针对珍稀的南美直嘴芦雀（*Limnoctites rectirostris*，灶鸟科Furnariidae）展开分布建模，该物种为沼泽生境专性类群，依赖多刺草本刺芹属（*Eryngium* spp.，伞形科Apiaceae）获取主要的食物与巢材资源。为此，我们首先对三种被认定为主要生物交互对象的刺芹属物种（*E. eburneum*、*E. horridum*与*E. pandanifolium*）开展分布建模，并将结果纳入直嘴芦雀的物种分布模型中。其次，我们采用动态范围框（dynamic range boxes）——一种用于量化n维超体积（n-dimensional hypervolumes）大小的创新方法，基于环境变量分析了直嘴芦雀与刺芹属物种间的生态位重叠情况。
相较于仅使用气候变量构建的模型，纳入生物交互作用后模型的预测性能得到显著提升。*E. eburneum*的气候适宜性与最湿季平均温度是模型中最重要的预测因子。与之相反，*E. horridum*与*E. pandanifolium*的预测效果欠佳，这表明直嘴芦雀对不同刺芹属物种的相对依赖程度存在差异。三种刺芹属物种的环境空间基本覆盖了直嘴芦雀的环境空间，但反之则不成立。
本研究结果表明，生物交互作用在大尺度下解释与预测珍稀鸟类分布方面发挥着重要作用；同时，对交互物种间生态位重叠的评估，可验证或加深我们对珍稀隐匿鸟类的个体生态学认知。我们倡导采用整合气候变量与生物交互作用的物种分布模型建模框架，以提升对认知匮乏鸟类物种的生态学认知水平。