The scale dependency of trait‐based tree neighborhood models

Abstract

Questions: We asked whether: 1) the strength of conspecific and heterospecific neighborhood crowding effects on focal tree survival and growth vary with neighborhood radii and 2) if the relative strength of the effect of neighborhood interactions on tree growth and survival varies with neighborhood scale? Location: Luquillo Forest Dynamics Plot, Puerto Rico. Methods:We used tree survival and growth data and included information on species-mean trait values related to several leaf traits, maximum height, seed mass and wood density. We incorporated a tree neighborhood modeling approach that uses an area around a focal tree with a specified radius, to describe the interactions between a focal tree and its neighbors. We constructed survival and growth models for each functional trait using a Bayesian approach, and varied the size of the radius from 5m to 30m, at 5m intervals. Results:The results suggested that the estimated effects of conspecific and heterospecific neighbors on tree performance do not vary based on the size of the neighborhood (5-30m), suggesting that the effects of conspecific and heterospecific neighbors on the performance of a focal tree likely do not vary substantially beyond a neighborhood radius of 5 meters in the Luquillo forest. In contrast, the estimated strength of the functional neighborhood (effect of neighbors based on their functional trait values) on tree performance was dependent on the neighborhood range. Our results also suggested that the effects of trait distances and trait hierarchies on tree survival and growth are acting simultaneously and at the same spatial scales. Conclusion:Findings from this study highlight the importance of spatial scale in community assembly processes, and specifically, a call for increased attention when selecting the radius that defines the neighborhood around a focal tree as the selected neighborhood radius influences the Accepted ArticleThis article is protected by copyright. All rights reserved community patterns discovered, and affects the conclusions about the drivers that control community assembly.

Light can modify density‐dependent seedling mortality in a temperate forest

“…Our results illustrate that woody seedlings may respond to both seedling and tree neighbours at different spatial scales, depending on the size class, growth form and species of the seedlings. This is consistent with current views that processes operating at plant neighbourhood scales play a fundamental role in structuring plant communities (Adler et al, 2018) and, accordingly, the size of plant neighbourhoods as well as the identity of plant neighbours are essential for understanding how plant performance is influenced by species interactions (Zambrano et al, 2020).…”

“…Because plants may differ in the strength and spatial scale of density-dependent effects (Zambrano et al, 2019;Zambrano et al, 2020), we used nine competing models, obtained for all combinations of three neighbourhood sizes (5-, 10-and 15-m radius) and three types of tree neighbours (conspecific, heterospecific and all), to select a neighbourhood size for each type of tree neighbour (Appendix S2 Figures S1-S2). In addition, we obtained three models with all tree neighbours (Appendix S2 Figures S3-S4).…”

“…We cannot, however, account for the causes of this scale dependence in our study because the zone of influence of seedling and tree neighbours on seedling mortality is likely to result from speciesspecific processes and life histories (Lasky et al, 2015;Detto et al, 2019;Zambrano et al, 2019;Zambrano et al, 2020). Seedling response to conspecific tree neighbours may be an outcome of multiple and often highly complex processes, such as pressure from different loads of host-specific enemies (Liang et al, 2016;Liu & He, 2019), intraspecific interactions (Adler et al, 2018) and mycorrhizal infections (Waud et al, 2016;Wulantuya et al, 2020), likely operating at scales both within and beyond seedling neighbourhoods.…”

“…As density-dependent effects have been traditionally studied using a fixed neighbourhood size to capture the interactions between seedlings and their neighbours (Comita et al, 2009;Bai et al, 2012;Johnson et al, 2017), such an approach does, in our opinion, not fully account for the fact that the processes underlying these interactions may operate at different spatial scales (Zambrano et al, 2019;Zambrano et al, 2020). For instance, the spatial scale of neighbourhoods over which seedlings experience density-dependent effects may differ among species (Queenborough et al, 2007), and likely also among plants of different sizes and growth forms.…”

“…We also used statistical neighbourhood models with variable neighbourhood size (Zambrano et al, 2020) to assess density-dependent effects (Piao et al, 2013;Johnson et al, 2014;Brown et al, 2020).…”