Spatial analyses for non‐overlapping objects with size variations and their application to coral communities
Spatial analyses for non‐overlapping objects with size variations and their application to coral communities
复制标题
具有尺寸变化的非重叠物体的空间分析及其在珊瑚群落中的应用
DOI：
10.1111/1365-2656.12193
复制
发表时间：
2013
影响因子：
4.8
通讯作者：
Y. Nozawa
中科院分区：
文献类型：
--
作者：
Muko S.;I.K. Shimatani;Y. Nozawa
个体的空间分布通常是通过将对象表示为无量纲点来分析的，其中空间统计是基于中心到中心的距离。然而，如果生物在没有重叠的情况下扩张，并表现出大小变化，比如覆盖珊瑚，那么物体间的间距对于发生相互作用的空间关联至关重要。我们引入了新的两两统计，使用物体之间的最小距离，并展示了它们在检查珊瑚群落数据时的效用。我们还计算了传统的点过程统计和基于网格的统计，以阐明每种空间统计方法的优点和局限性。为简单起见，在这些演示中，珊瑚群落用圆盘来表示。关注短距离效应，最小距离的使用表明，几乎所有珊瑚属都聚集在1-25厘米的尺度上。然而，当破碎的菌落（株）被视为一个基因时，基因水平的分析表明弱聚集或没有聚集，这表明大多数珊瑚是随机分布的，碎片化是菌落聚集的主要原因。相比之下，点过程统计显示更大的聚集尺度，可能是因为中心到中心的距离包括了群体间距和群体大小（半径）。基于网格的统计能够量化菌落的斑块（聚集）规模，但规模受到菌落大小的强烈影响。我们的方法定量地显示了侵略性属和竞争性弱属之间的排斥效应，而基于网格的统计（协方差函数）也显示了排斥，尽管统计数据显示的空间尺度不能直接解释生态意义。最小距离的使用与先前提出的空间统计数据一起帮助我们扩展了对大小和相关特定尺度不同的非重叠物体的空间模式的理解。
Spatial distributions of individuals are conventionally analysed by representing objects as dimensionless points, in which spatial statistics are based on centre‐to‐centre distances.However, if organisms expand without overlapping and show size variations, such as is the case for encrusting corals, interobject spacing is crucial for spatial associations where interactions occur.We introduced new pairwise statistics using minimum distances between objects and demonstrated their utility when examining encrusting coral community data. We also calculated the conventional point process statistics and the grid‐based statistics to clarify the advantages and limitations of each spatial statistical method. For simplicity, coral colonies were approximated by disks in these demonstrations.Focusing on short‐distance effects, the use of minimum distances revealed that almost all coral genera were aggregated at a scale of 1–25 cm. However, when fragmented colonies (ramets) were treated as a genet, a genet‐level analysis indicated weak or no aggregation, suggesting that most corals were randomly distributed and that fragmentation was the primary cause of colony aggregations. In contrast, point process statistics showed larger aggregation scales, presumably because centre‐to‐centre distances included both intercolony spacing and colony sizes (radius). The grid‐based statistics were able to quantify the patch (aggregation) scale of colonies, but the scale was strongly affected by the colony size.Our approach quantitatively showed repulsive effects between an aggressive genus and a competitively weak genus, while the grid‐based statistics (covariance function) also showed repulsion although the spatial scale indicated from the statistics was not directly interpretable in terms of ecological meaning.The use of minimum distances together with previously proposed spatial statistics helped us to extend our understanding of the spatial patterns of nonoverlapping objects that vary in size and the associated specific scales.