School of Agriculture, Policy and Development, University of Reading, Reading, UK.
Institute of Forestry & Conservation, University of Toronto, Toronto, Ontario, Canada.
Ecol Appl. 2023 Jun;33(4):e2838. doi: 10.1002/eap.2838. Epub 2023 Apr 26.
Meta-analyses have become a valuable tool with which to synthesize effects across studies, but in ecology and evolution, they are often characterized by high heterogeneity, where effect sizes vary between studies. Much of this heterogeneity can be attributed to species-specific differences in responses to predictor variables. Here, we aimed to incorporate a novel trait-based approach to explain species-specific differences in a meta-analysis by testing the ability of morphological traits to explain why the effectiveness of flight-intercept trap design varies according to beetle species, a critical issue in forest pest management. An existing morphological trait database for forest beetles was supplemented, providing trait data for 97 species, while data from a previous meta-analysis on capture rates of bark or woodboring beetles according to different trap designs were updated. We combined these sources by including nine morphological traits as moderators in meta-analysis models, for five different components of trap design. Traits were selected based on theoretical hypotheses relating to beetle movement, maneuverability, and sensory perception. We compared the performance of morphological traits as moderators versus guild, taxonomic family, and null meta-analysis models. Morphological traits for the effect of trap type (panel vs. multiple-funnel) on beetle capture rates improved model fit (AIC ), reduced within-study variance (σ ), and explained more variation (McFadden's pseudo-R ) compared with null, guild, and taxonomic family models. For example, morphological trait models explained 10% more of the variance (pseudo-R ) when compared with a null model. However, using traits was less informative to explain how detailed elements of trap design such as surface treatment and color influence capture rates. The reduction of within-study variance when accounting for morphological traits demonstrates their potential value for explaining species-specific differences. Morphological traits associated with flight efficiency, maneuverability, and eye size were particularly informative for explaining the effectiveness of trap type. This could lead to improved predictability of optimal trap design according to species. Therefore, morphological traits could be a valuable tool for understanding species-specific differences in community ecology, but other causes of heterogeneity across studies, such as forest type and structure, require further investigation.
元分析已成为综合研究结果的一种有价值的工具，但在生态学和进化生物学中，它们通常具有高度的异质性，即研究之间的效应大小存在差异。这种异质性的很大一部分可以归因于物种对预测变量的反应存在特异性差异。在这里，我们旨在通过测试形态特征解释飞行拦截陷阱设计有效性因甲虫物种而异的能力，将一种新的基于特征的方法纳入元分析中，以解释物种特异性差异，这是森林害虫管理中的一个关键问题。我们补充了一个现有的森林甲虫形态特征数据库，提供了 97 个物种的特征数据，同时更新了之前关于不同陷阱设计对树皮或木质甲虫捕获率的元分析数据。我们通过将九个形态特征作为调节因子纳入元分析模型中，将这些来源结合起来，针对五种不同的陷阱设计组件。特征是根据与甲虫运动、操纵和感官感知相关的理论假设选择的。我们将形态特征与 guild、分类家族和 null 元分析模型进行了比较。形态特征对于陷阱类型（面板与多漏斗）对甲虫捕获率的影响，改善了模型拟合（AIC ），减少了研究内方差（σ），并解释了更多的变异（麦克法登的伪 R ），与 null、guild 和分类家族模型相比。例如，与 null 模型相比，形态特征模型解释了 10%更多的变异（伪 R ）。然而，使用特征来解释陷阱设计的详细元素（如表面处理和颜色）如何影响捕获率的信息较少。在考虑形态特征时，研究内方差的减少证明了它们在解释物种特异性差异方面的潜在价值。与飞行效率、操纵性和眼睛大小相关的形态特征对于解释陷阱类型的有效性特别有帮助。这可以提高根据物种优化陷阱设计的可预测性。因此，形态特征可能是理解群落生态学中物种特异性差异的有价值的工具，但研究之间的其他异质性原因，如森林类型和结构，需要进一步调查。
