Data_Sheet_1_The Shape of Species Abundance Distributions Across Spatial Scales.docx

Species abundance distributions (SADs) describe community structure and are a key component of biodiversity theory and research. Although different distributions have been proposed to represent SADs at different scales, a systematic empirical assessment of how SAD shape varies across wide scale gradients is lacking. Here, we examined 11 empirical large-scale datasets for a wide range of taxa and used maximum likelihood methods to compare the fit of the logseries, lognormal, and multimodal (i.e., with multiple modes of abundance) models to SADs across a scale gradient spanning several orders of magnitude. Overall, there was a higher prevalence of multimodality for larger spatial extents, whereas the logseries was exclusively selected as best fit for smaller areas. For many communities the shape of the SAD at the largest spatial extent (either lognormal or multimodal) was conserved across the scale gradient, despite steep declines in area and taxonomic diversity sampled. Additionally, SAD shape was affected by species richness, but we did not detect a systematic effect of the total number of individuals. Our results reveal clear departures from the predictions of two major macroecological theories of biodiversity for SAD shape. Specifically, neither the Neutral Theory of Biodiversity (NTB) nor the Maximum Entropy Theory of Ecology (METE) are able to accommodate the variability in SAD shape we encountered. This is highlighted by the inadequacy of the logseries distribution at larger scales, contrary to predictions of the NTB, and by departures from METE expectation across scales. Importantly, neither theory accounts for multiple modes in SADs. We suggest our results are underpinned by both inter- and intraspecific spatial aggregation patterns, highlighting the importance of spatial distributions as determinants of biodiversity patterns. Critical developments for macroecological biodiversity theories remain in incorporating the effect of spatial scale, ecological heterogeneity and spatial aggregation patterns in determining SAD shape.

物种多度分布（Species Abundance Distributions，SADs）用于描述群落结构，是生物多样性理论与研究的核心组成部分。尽管已有不同分布模型被提出，用以表征不同尺度下的SADs，但目前仍缺乏针对SAD形态如何随大范围尺度梯度变化的系统性实证评估。本研究针对广泛类群的11套大型实证数据集展开分析，采用最大似然法，在跨越数个数量级的尺度梯度上，对比对数级数分布、对数正态分布以及多峰（即存在多个多度峰值）模型对SADs的拟合效果。整体而言，空间范围越大，多峰分布的占比越高；而对数级数分布仅在小型区域中被选为最优拟合模型。尽管采样区域面积与类群多样性均出现大幅下降，但多数群落的最大空间尺度下的SAD形态（对数正态或多峰分布）在整个尺度梯度中保持稳定。此外，SAD形态会受到物种丰富度的影响，但未检测到个体总数对其产生系统性影响。本研究结果表明，现有两大主流生物多样性宏观生态学理论对SAD形态的预测与实际情况存在显著偏离。具体而言，无论是生物多样性中性理论（Neutral Theory of Biodiversity，NTB）还是生态最大熵理论（Maximum Entropy Theory of Ecology，METE），均无法解释本研究中观测到的SAD形态变异。这一点体现在两方面：一是对数级数分布在大尺度下的拟合效果不佳，与NTB的预测相悖；二是各尺度下的SAD形态均偏离METE的理论预期。重要的是，这两种理论均未考虑SADs中的多峰分布现象。我们认为，本研究结果由种间和种内空间聚集模式共同支撑，凸显了空间分布作为生物多样性格局决定因素的重要性。未来生物多样性宏观生态学理论的关键发展方向，在于将空间尺度、生态异质性以及空间聚集模式对SAD形态的影响纳入考量。