Forest structure tracks Pollino plant composition nearly twice as closely as topography

The three shares sum to 67.6 percent. That leaves 32.4 percent of the variation in vascular plant composition unattributed to either set of variables ^[1]. Structure's unique contribution is about 1.8 times topography's ^[2]. The joint share, the 31.3 percent the analysis cannot hand to one side or the other, is bigger than either unique share, and exceeds structure's by 7.8 points ^[4].

Some of that overlap was inevitable, because terrain shapes the stands. Topographic variables accounted for about 18.8 percent of the variation in structural attributes across the full gradient, and the authors report that topography does influence forest structure while explaining only part of its variation ^[9][15].

The design was built to separate the two. "By combining measurements of living trees, large-diameter trees, canopy structure and deadwood with floristic and topographic data, we were able to separate the contribution of forest structure from that of topography and examine how structural differences are associated with vascular plant communities," said Danilo Travascia of the University of Basilicata, the first author ^[13].

Nine candidate old-growth sites were surveyed and the final analyses used 142 subplots at eight of them, roughly 18 subplots per site ^[3][4][3]. One surveyed site did not reach the final analyses, and the phys.org account does not say why ^[5]. Living trees, regeneration, standing and downed deadwood, coarse woody debris and decay stages were recorded alongside the vascular flora and the topographic variables, yielding 24 structural variables ^[6][5]. Structural complexity was summarised in a Structural Heterogeneity Index, and the team used mixed-effects models with canonical correspondence analysis, principal coordinate analysis and variation partitioning ^[7]. The floristic results were also put through sampling-completeness and sensitivity checks ^[8].

"By surveying living trees, regeneration, canopy attributes, deadwood and the entire vascular flora at the same spatial scale, we could analyze each stand as an integrated ecological system," said Francesco Ripullone, the senior author ^[14].

Nothing was manipulated. The team measured the stands as they stood ^[6], and all the sites were candidate old-growth forests ^[3], so the contrast runs among mature stands rather than between mature and managed ones. The reported percentages describe composition, meaning which species occur together, and not a count of species per plot ^[10]. The attributes that travelled most closely with the flora were large-tree abundance, canopy cover and complexity, tree density and heterogeneity, tree species richness, coarse woody debris and the diversity of deadwood decay stages ^[12]. Those features, the authors write, create a mosaic of light conditions, microclimates, substrates and microhabitats for plants ^[16].

So a manager deciding what to leave inside a Pollino stand gets a named list of correlates from this work ^[12]. A manager deciding where to draw the next reserve boundary gets less, since terrain and structure share 31.3 points of the explained variation and the survey covered only old-growth candidates ^[11][3].