Insect herbivory and elemental fluxes
The mean leaf area lost to BIH across all sites was 1.6±0.33%, while 10.0±1.6% of the leaves sampled across all sites were subjected to some degree of insect herbivory (Table 1). The proportional leaf area lost to herbivory was positively correlated with the relative position above the valley bottom (p<0.001), annual solar radiation (p=0.028), and leaf N-content (p=0.028), yet showed no link to temperature, absolute elevation, or leaf content of condensed tannins (Table 2).
The overall mean elemental canopy loss to herbivores was estimated to 0.25±0.048 g C m-2 yr-1, 0.017±0.0033 g N m-2 yr-1, and 0.0017±0.00032 g P m-2 yr-1 (Table 1). The mean leaf production across sites was 47±5 g dry matter m-2 yr-1, corresponding to 21±2 g C m-2 yr-1, 0.6±0.08 g N m-2 yr-1, and 0.09±0.011 g P m-2 yr-1 (Table 1). The estimated proportion of nutrients resorbed from the birch leaves towards the end of the season was 56±2.8% for N and 30±4.1% for P (Table 1). The resulting proportion of the annual flux from the canopy through insect deposits make up 1.5±0.3% of the total C flux (litter+herbivore input), while the proportion of N and P were slightly higher (3.6±0.7% and 2.4±0.5%, respectively) due to foliar nutrient resorption during senescence. The annual proportion of canopy-to-soil N and P fluxes caused by herbivory were positively correlated with herbivory level (N and P: p<0.001) and resorption (N: p=0.002, P: p=0.024). Moreover, the proportional insect mediated flux of N was negatively correlated with leaf production (N: p=0.005). There was no significant link to leaf N or P concentrations (Table S2).