Table_1_Interpreting spatially explicit variation in dietary proxies through species distribution modeling reveals foraging preferences of mammoth (Mammuthus) and American mastodon (Mammut americanum).XLSX

IntroductionThe end Pleistocene was a time of considerable ecological upheaval. Recent work has explored the megafauna extinction’s role in altering ecosystem processes. Analyses of functional traits withing communities reveal hidden consequences of the megafauna extinction beyond declines in taxonomic diversity. Functional diversity analyses offer new insight into our understanding of past ecosystems and may even inform future rewilding efforts. However, the utility of functional diversity may be hampered by the use of discrete, taxon-level functional traits, such as dietary categories, that mask variation in functional diversity over space and time.
MethodsWe present an approach in which species distribution modeling, in Maxent, provides context for interpreting variation in two widely used proxies for diet among fossil taxa: stable isotope analysis and dental microwear texture analysis. We apply this approach to two ecologically distinct taxa, the American mastodon (Mammut americanum) and mammoths (Mammuthus) and investigate their resource use over space and time from the last glacial maximum to the end Pleistocene (25–11.7 thousand years before present).
ResultsMammoth dietary behavior varies by context across their geographic distribution, despite possessing evolutionary adaptations that facilitate grazing. Mammoths exhibit a preference for grazing where species distribution modeling predicts the highest likelihood of occurrence but engage in more mixed-feeding outside of core likelihood areas. In contrast, dietary preferences for mastodon are less resolved and our analyses were unable to identify significant differences in diet across their distribution.
DiscussionThe ecological roles of some species are context specific and need to be critically evaluated when planning for management of reintroductions or introducing novel species to restore lost ecological function.

引言
更新世末期是一场规模显著的生态剧变时期。近期研究已围绕巨型动物群灭绝对生态系统过程的影响展开了诸多探索。针对群落内功能性状的分析揭示了巨型动物群灭绝所带来的、远超分类多样性下降之外的隐性生态后果。功能多样性分析可为我们理解古生态系统提供全新视角，甚至可为未来的再野化工作提供参考依据。然而，采用离散的、分类单元级别的功能性状（如饮食类别）开展的研究，可能会限制功能多样性分析的应用效用——这类性状会掩盖功能多样性在时空维度上的变异。

研究方法
本文提出一种研究方法：借助最大熵物种分布模型（Maxent），为解读化石分类单元中两种常用的饮食替代指标——稳定同位素分析与牙齿微磨耗纹理分析——的变异情况提供情境参考。我们将该方法应用于两个生态特征迥异的类群：美洲乳齿象（*Mammut americanum*）与猛犸象（*Mammuthus*），并探究了从末次盛冰期至更新世末期（距今2.5万至1.17万年）期间二者资源利用模式的时空变化。

研究结果
尽管猛犸象演化出了利于放牧的适应性特征，但其饮食行为在不同地理分布区域内存在显著的情境依赖性差异。在最大熵物种分布模型预测的高适宜性分布区域内，猛犸象倾向于取食草本植物；而在核心适宜区域之外，则更多采取混合取食策略。与之形成对比的是，美洲乳齿象的饮食偏好尚未得到明确解析，我们的分析未能在其分布范围内检测到饮食模式的显著差异。

讨论
部分物种的生态功能具有情境特异性，在规划物种再引入管理或引入外来物种以恢复丧失的生态功能时，需对其进行审慎评估。