Fragmentation and matrix contrast favor understory plants through negative cascading effects on a strong competitor palm

Abstract

Understanding the patterns and processes driving biodiversity maintenance in fragmented tropical forests is urgently needed for conservation planning, especially in species-rich forest reserves. Of particular concern are the effects that habitat modifications at the landscape scale may have on forest regeneration and ecosystem functioning: a topic that has received limited attention. Here, we assessed the effects of landscape structure (i.e., forest cover, open area matrices, forest fragmentation, and mean inter-patch isolation distance) on understory plant assemblages in the Los Tuxtlas Biosphere Reserve, Mexico. Previous studies suggest that the demographic burst of the strong competitor palm Astrocaryum mexicanum in the core area of this reserve limits plant recruitment and imperils biodiversity conservation within this protected area. Yet, the local and landscape predictors of this palm, and its impact on tree recruitment at a regional scale are unknown. Thus, we used structural equation modeling to assess the direct and cascading effects of landscape structure on stem and species density in the understory of 20 forest sites distributed across this biodiversity hotspot. Indirect paths included the effect of landscape structure on tree basal area (a proxy of local disturbance), and the effects of these variables on A. mexicanum. Density of A. mexicanum mainly increased with decreasing both fragmentation and open areas in the matrix (matrix contrast, hereafter), and such an increase in palm density negatively affected stem and species density in the understory. The negative direct effect of matrix contrast on stem density was overridden by the indirect positive effects (i.e., through negative cascading effects on A. mexicanum), resulting in a weak effect of matrix contrast on stem density. These findings suggest that dispersal limitation and negative edge effects in more fragmented landscapes dominated by open areas prevent the proliferation of this palm species, enhancing the diversity and abundance of understory trees. This "positive" news adds to an increasing line of evidence suggesting that fragmentation may have some positive effects on biodiversity, in this case by preventing the proliferation of species that can jeopardize biodiversity conservation within tropical reserves.

Landscape structure shapes tree seedlings’ diversity at multiple spatial scales in a fragmented tropical rainforest

“…These findings support the assumption the matrix drives functional connectivity of seed dispersers at smaller spatial scales [16]. It is noteworthy that the observed scale of effect of EC (800-m) is similar to those of matrix contrast observed for understory plants (797.88-m) and seed rain abundance (800±128 m) in other HMTLs of Mexico [16,35].…”

“…The overall classification accuracy was 79%. We then calculated six landscape metrics (Table 1) relevant to the diversity of seed and plant communities [16,35]. The composition metrics included the percentages of old-growth forest (OGF) and secondary forest (SF) covers, whereas the configuration metrics were the density (PD) and aggregation (AI) of OGF patches.…”

“…We selected the mean Euclidean-nearest neighbor distance between patches (ENN) for estimating structural connectivity and the percentage of contrasting edges (EC) as a measure of functional connectivity. [16,19,27,35,59–63]…”

“…We estimated these metrics within 13 circular buffers (of a 300 to1500-m radius, at 100 m intervals) from the center of each focal forest patch (Figure 1c). The number and range of the buffers’ sizes comprise the reported scale of effect for a variety of patterns in understory vegetation, seed rain, and bird and mammal communities in HMTLs [27,35,60,62]. Although the overlapping between buffers of nearby sampling points increased at larger spatial extents, the degree of spatial autocorrelation in the model residuals is not necessarily associated with the extent of landscape overlap but rather with the proximity between sampling sites [75,76].…”

“…This avoids overlooking the true scale of effect if assessments were performed using very few scales within narrow ranges [29,34]. Only a handful of studies have employed multi-scale approaches for plant communities within HMTLs [16,35], and to our knowledge, there is a lack of studies on seedling communities. In this study, we employed a multi-scale approach to assess the contribution of landscape composition (OGF and SF amount), configuration (the patches’ aggregation and fragmentation) and connectivity (structural and functional) on α- and β-diversity of tree seedlings within rainforest patches.…”