Data from: Localized hotspots drive continental geography of abnormal amphibians on U.S. wildlife refuges

Amphibians with missing, misshapen, and extra limbs have garnered public and scientific attention for two decades, yet the extent of the phenomenon remains poorly understood. Despite progress in identifying the causes of abnormalities in some regions, a lack of knowledge about their broader spatial distribution and temporal dynamics has hindered efforts to understand their implications for amphibian population declines and environmental quality. To address this data gap, we conducted a nationwide, 10-year assessment of 62,947 amphibians on U.S. National Wildlife Refuges. Analysis of a core dataset of 48,081 individuals revealed that consistent with expected background frequencies, an average of 2% were abnormal, but abnormalities exhibited marked spatial variation with a maximum prevalence of 40%. Variance partitioning analysis demonstrated that factors associated with space (rather than species or year sampled) captured 97% of the variation in abnormalities, and the amount of partitioned variance decreased with increasing spatial scale (from site to refuge to region). Consistent with this, abnormalities occurred in local to regional hotspots, clustering at scales of tens to hundreds of kilometers. We detected such hotspot clusters of high-abnormality sites in the Mississippi River Valley, California, and Alaska. Abnormality frequency was more variable within than outside of hotspot clusters. This is consistent with dynamic phenomena such as disturbance or natural enemies (pathogens or predators), whereas similarity of abnormality frequencies at scales of tens to hundreds of kilometers suggests involvement of factors that are spatially consistent at a regional scale. Our characterization of the spatial and temporal variation inherent in continent-wide amphibian abnormalities demonstrates the disproportionate contribution of local factors in predicting hotspots, and the episodic nature of their occurrence.近二十年来，出现肢体缺失、畸形或多肢异常的两栖动物，已持续引发公众与科学界的广泛关注，但该现象的整体发生规模仍未得到充分认知。尽管学界已在部分地区明确了此类畸形发生的诱因，但由于对其更广泛的空间分布与时间动态缺乏系统认知，目前仍难以厘清其对两栖动物种群衰退及环境质量的潜在影响。为填补这一数据空白，我们针对美国国家野生动物保护区（U.S. National Wildlife Refuges）内的62947只两栖动物开展了为期十年的全国性评估。对包含48081个个体的核心数据集（core dataset）的分析显示，与预期的背景频率相符，平均有2%的个体存在畸形异常，但异常发生情况呈现显著的空间异质性，最高异常率可达40%。方差分解分析（variance partitioning analysis）结果表明，与空间相关的因素（而非物种或采样年份）可解释97%的畸形异常变异，且分解得到的变异占比随空间尺度扩大（从样点到保护区再到区域）而逐渐降低。与此一致，畸形异常呈现从局地到区域尺度的热点聚集特征，在数十至数百公里的空间尺度上形成集群。我们在密西西比河流域、加利福尼亚州与阿拉斯加州均发现了此类高异常率站点的热点集群。热点集群内部的异常频率波动幅度显著高于集群外部。这一结果与干扰或天敌（病原体或捕食者）等动态致因的作用模式相符；而在数十至数百公里尺度上异常频率的相似性，则表明存在区域尺度上空间一致性的驱动因素。我们对全美范围内两栖动物畸形异常固有空间与时间变异特征的刻画，证实了局地因素在预测畸形热点时具有远超预期的重要性，同时也表明此类畸形热点的发生具有偶发性特征。