Factors affecting tree crown allometries and consequences on forest structure

Tree crown architecture is crucial for light capture, it determines the competitive advantage over neighbouring trees and related to the demographic process of forests. Geographic variation of crown architecture is impressive: from flat-top crown in tropical regions to narrow-deep in boreal regions. However, the assessment of crown traits across geographic areas is still lacking. In this research, I would like to figure out the relationship of crown properties with size by the simply effective power law method, namely allometry. With the aim of quantifying the relationship at broad scales, we collected data across latitudes (40.65° S to 67.95° N), and data across elevations (0m - 2150m a.s.l.) in a narrow latitude ranges (45° - 46° N) by public dataset and our own measurements mixed with different species. We also collected data of branch properties (length, diameter, geometry etc) in a temperate forest to test whether branch allometries and branch-size distribution parallel what is observed in individual trees and in the whole forest. Our aim was to answer to the following three main questions: 1) Does the tree crown geometry change with latitude? 2) Does the tree crown geometry change with elevation? 3) Do geometry and distribution of branches mimic, respectively, tree crown geometry and forest structure? Results showed that the relationship of the crown radius (i.e the growth in crown width) vs. tree height was clearly latitude dependent with decreased scaling exponents from ~1 to ~0.5 moving from inter-tropical to boreal areas irrespectively of the phyletic effects. This suggests that trees prioritize the height growth comparing with lateral growth as latitude increases and, as a consequence, that crowns become more and more elongated with growth at higher latitudes. However, the crow length scaled isometrically (≈1) with tree height in all latitudes, showing that the relative length of stem with living branches is maintained nearly constant as trees grow taller. Thus the scaling of crown volume with tree height showed the same geographical pattern as crown radius, and the scaling exponent decreased from ~3 to ~2 from the tropics to boreal forests. When tree crown geometry is compared at the same latitude but along an altitudinal gradient no significant trends of crown shape were detected thus reinforcing the idea that solar elevation angle might be the causal explanation for thinner crown at high latitudes. Relatively to branch geometry, results showed that, in conifers, the leaves were essentially spread in 2 dimensions (single branches fill a planar surface), but the branch leaf area scaled vs. branch length with the same exponent of tree crown volume vs. tree height, namely ~2.3-2.5 in temperate forests. However the strategies for reaching the same leaf accumulation are different. Secondary branches grow in width much more (i.e. with exponent >1 vs. branch length) than primary branches in a tree. Branch size distribution in a tree, approximates a power law with exponent driven by the scaling exponent of branch area vs. branch length, and, remarkably, behaved similar to the tree size distribution in forests that is derived by the scaling of crown volume at individual level. Above all, our result revealed the crown allometries are not universal, which crown radius and crown volume showed strong correlations with latitude but not correlated with elevation. Given the relationship between crown structure and tree-size distribution, this framework provides the causal explanation for the different forest structures observed across the globe. Besides, the consistent allometric exponent between tree crown volume and branch volume with size suggests the same physiological needs for surviving in a given area, which implies the entire forest is constructed by a hierarchical system characterized by the same rate of energy utilization, from tree to different branch orders.