Trait multifunctionality sets individual traits sub-optimal for specific functions \cite{Sack_2019}. Whole-tree size structural traits are typically multifunctional \cite{Poorter_2003} and may be both polygenic and pleiotropic \cite{Farnsworth_1995}, resulting from hierarchical, emergent consolidation of lower-level traits \cite{Charles_Dominique_2012}. As a result, high co-linearity is usually found among the tree size traits \cite{Martin_Ducup_2020,Verbeeck_2019}, making the contributions to individual functions hard to disentangle \cite{Messier_2017,Sack_2019}. Nonetheless, a division of the size traits into vertical and horizontal extents was proposed \cite{Seidel_2019}, reflecting the predominant linkages to functions of space exploration and light harvest, respectively \cite{Hall__2001}. A better insight into the structure-function network in trees may be provided by including the size-controlled relative traits, or architectural proportions \cite{Iida_2011,MacFarlane_2017}, with both a higher level of trait orthogonality and specific trade-offs (negatively correlated traits) within the major trait dimensions \cite{Martin_Ducup_2020,K_dra_2022}. Using this approach, it was revealed that trees may amplify specific functions along ontogeny and increasing size, such as for light harvest and reproduction, resulting in architecture convergence with increasing canopy position in large-statured trees \cite{Martin_Ducup_2020}. However, architecture divergence \cite{Loubota_Panzou_2018} was also reported when both large- and small-statured trees were included, suggesting two or more viable architectural trait combinations. The architecture convergence-divergence hypotheses require further testing \cite{Martin_Ducup_2020,Loubota_Panzou_2018}. Particularly, of interest are the intraspecific architecture-function trajectories in the small-statured species of mixed life-history strategy, while it appears that such species have a predominant effect on the community-level structural diversity \cite{Iida_2011,Loubota_Panzou_2018}, exhibiting more than a single architectural optimum depending on local environment and type of stress \cite{Sack_2019}.
Rowan Sorbus aucuparia L. (Rosaceae family) is a widely distributed broad-leaved deciduous tree species, occurring across a diverse range of European landscapes, canopy openness gradients, and from the lowlands up to subalpine scrub land at 2000 m a.s.l. \cite{2000}. Near the upper limit, rowan maintains a tree growth-form, above other tree species \cite{Barclay_1982}. It is the sole understory tree in the Carpathian Norway spruce (Picea abies) subalpine stands \cite{_ywiec_2007}. These forests are subject to dynamic changes \cite{_ywiec_2012,Holeksa_2017} and despite a monospecific canopy, the rowans have very diverse exposures to different stressors, such as low light and high wind \cite{McIntire_2016}. The large fleshy-fruit production in such rowan trees was recently confirmed related positively to both stem girth \cite{Bogdziewicz_2020} and light availability \cite{Kondrat_2024}, but with a very large variation remaining unexplained. Specifically, when fruit production was standardized by tree size, its both highest and lowest values were recorded at intermediate light availability \cite{Kondrat_2024}, suggesting more complex structure-environment-fecundity patterns \cite{Davi_2016,Innes_1994}. Particularly, the effects of crown traits may be non-monotonic \cite{Davi_2016}, driven by the growth-reproduction trade-offs \cite{Innes_1994}.
Here, we assume that the rowan architecture varies along the size \cite{Seidel_2019} and the proportions \cite{Martin_Ducup_2020,K_dra_2022} structural trait dimensions (Fig. \ref{435780}), with horizontally-developed trees being stouter, but higher trees not necessary having longer crowns \cite{Pretzsch_2012}. The two extreme trait combinations (Fig. \ref{435780}; B,C) are expected rare in the subalpine rowans,  disregarding canopy openness. Namely: stout trees with full and costly crowns (Fig. \ref{435780}; B), limited by harsh conditions \cite{Szwagrzyk_2015}; and slender, but top-heavy trees (Fig. \ref{435780}; C), vulnerable to mechanical failure \cite{SPERRY_2007}. While the two intermediate trait combinations (Fig. \ref{435780}; A,D) are expected to be viable here, but with contrasting reproductive effects, depending on locally prevailing stresses, substantiated by varying canopy openness \cite{Westerband_2015}. The long and narrow crowns are linked to increased hydraulic safety and water use efficiency, due to shorter crown-root system path lengths \cite{Nunes_2023,Sack_2019}, and may be optimized for an intermediate reproductive performance at small gaps (Fig. \ref{435780}; A1). In more open situations (Fig. \ref{435780}; A2), when such slender trees are suddenly exposed to more sunlight and wind, there may be a shift towards the stem secondary growth \cite{Bonnesoeur_2016} and the crown widening \cite{Szwagrzyk_2015}, at the cost of reproduction; alternatively, while species with compound leaves may lower the effective crown sail area and thus mechanical stress from wind drag \cite{Shenkin_2020}, such trees might also acclimate to higher wind stress by increasing crown transparency, keeping the slender form reproductive. In contrast, wide and short crowns are often considered optimized for survival only (Fig. \ref{435780}; D1), in stunted individuals \cite{Givnish_1988,PICKETT_1980}, but may be very reproductive in the open (Fig. \ref{435780}; D2), with thicker stems providing support and better storage capacity \cite{McNeil_2023,Sack_2019}, crucial for the costly fruit production in rowan \cite{h1999,i1990}. Overall, we test the hypotheses of interacting crown architecture and canopy openness effects on subalpine rowan fecundity, to aid in functionally meaningful interpretation of tree architecture. The specific objectives included: (i) evaluation of rowan structural diversity in the harsh subalpine climate; (ii) identification of the structural trade-offs governing the functional performance of rowans; (iii) quantification of tree architecture-environment-fecundity trajectories. We overcome the difficulties of acquiring tree architecture in montane terrain by using a light-weight photogrammetric equipment \cite{K_dra_2019}.