Estimating species richness of pitfall catches by non-parametric estimators
Estimating species richness of pitfall catches by non-parametric estimators
复制标题
通过非参数估计器估计陷阱渔获物的物种丰富度
DOI：
10.1078/0031-4056-00117
复制
发表时间：
2002
期刊：
影响因子：
2.3
通讯作者：
U. Brose
中科院分区：
文献类型：
--
作者：
U. Brose
摘要由于新的土地利用系统的加速应用，农业景观研究面临的挑战是提供其可持续性的结果。生物多样性的研究，进行了最少的抽样程序已越来越受欢迎。然而，关于昆虫，在这些研究中观察到的物种数量低估了真正的物种丰富度。在这些研究中，非参数估计可能提供一个更可靠的工具，以获得有关的真实物种丰富度的信息。本文的目的是评估如何准确和精确地估计物种丰富度的研究与陷阱样本与高采样强度的研究。这表明在收集昆虫群落物种丰富度数据方面可以节省多少精力。步甲甲虫的物种丰富度在10个临时湿地在东德农业景观进行了调查，每个站点的9个陷阱陷阱。物种丰富度的估计计算的情况下，每个站点有四个，五个或六个陷阱陷阱。在研究中，观察到的物种数和非参数估计，Chao2，Bootstrap，Jacknife1和Jacknife2，估计物种丰富度在每个场景中，分别。将这些假设情景中的估计数与基于整个数据集的估计数进行了比较。虽然Chao2是最准确和精确的估计量，但Bootstrap估计量的表现仅略好于观察到的物种数量。Jacknife1和Jacknife2表现居中。只有Chao2的表现比观察到的物种数更精确。总之，在小生境中进行的研究中，样本数量可能会减少到每个站点五个或六个陷阱。非参数估计提供了可能性，以减少采样工作，以及进行物种丰富度的研究，以获得更可靠的信息，物种丰富度重新分析研究已经进行。
Summary In consequence of accelerated applications of new land-use systems, agricultural landscape research is challenged to provide results on their sustainability. Studies on biodiversity that are carried out with minimal sampling programs have become increasingly popular. However, concerning insects, the observed number of species in these studies underestimates the true species richness. In such studies, the non-parametric estimators might provide a more reliable tool to gain information about the true species richness. The objective of this paper is to evaluate how accurately and precisely estimates of species richness in studies with few pitfall samples correlate with those of studies with higher sampling intensity. This indicates how much effort could be saved in collecting data on species richness of insect communities. The species richness of carabid beetles at ten temporary wetlands in the East-German agricultural landscape was surveyed by nine pitfall traps per site. Estimates of species richness were calculated in scenarios with four, five or six pitfall traps per site. The observed number of species and the non-parametric estimators, Chao2, Bootstrap, Jacknife1 and Jacknife2, were included in the study to estimate species richness in each scenario, respectively. The estimates in the scenarios were compared with estimations based on the whole data set. While Chao2 was the most accurate and precise estimator, the Bootstrap estimator performed only slightly better than the number of observed species. Jacknife1 and 2 performed intermediately. Only Chao2 performed more precisely than the observed number of species. In conclusion, the number of samples might be reduced to five or six pitfall traps per site in studies carried out in small habitats with minimal sampling programs. The non-parametric estimators provide possibilities to carry out studies on species richness with reduced sampling efforts as well as, to gain more reliable information on species richness re-analyzing studies already carried out.