Permanent site characteristics exert a larger influence than atmospheric conditions on leaf mass, foliar nutrients and ultimately aboveground biomass productivity of Salix miyabeana 'SX67'

Salix species are widely used for wood production, but the interactive effects of soil types and atmospheric conditions on Salix foliar nutrients and aboveground biomass productivity have not yet been elucidated. The objectives of this study were fourfold: (1) to assess the variation in foliar nutrients and leaf mass of Salix miyabeana 'SX67' grown as short rotation coppice (SRC) across three growing seasons and nine locations in Quebec with different permanent site characteristics and atmospheric conditions, (2) to test if atmospheric conditions and permanent site characteristics could explain the variation in foliar nutrients and leaf mass, (3) to develop models that consider foliar traits and nutrient interactions to produce more robust predictions of annual aboveground biomass yields, and (4) to compare nutritional requirements of 'SX67' to other Salix cultivars used for SRC. Leaf samples were collected over three growing seasons at all sites. For each site, atmospheric conditions were simulated and foliar nutrient levels were measured to perform centered log ratio (clr) transformations for each foliar nutrient. This approach considered foliar nutrient interactions and dealt with clr scores as linearly independent. The clr scores were more largely influenced by permanent site characteristics than by atmospheric conditions, despite large variations in degree-days. However, some foliar nutrients and leaf mass were linearly related to atmospheric variables within sites. Strong relationships between annual aboveground biomass yields and leaf mass were computed (e.g. adjusted R2 = 0.62), likely due to a proportional allocation between foliage and wood. Although significant linear relationships between clr scores (i.e. N, Ca and Mn) and annual aboveground biomass yields were detected, yields were more robustly explained non-linearly by thresholds (i.e. N, Ca and P) (e.g. R2 = 0.85), likely due to permanent characteristics specific to each of the sites and climatic limitations during the growing seasons studied. The thresholds detected by non-linear models suggested high N and P use efficiencies and a large Ca requirement of 'SX67'.