Effects of drying and connectivity on instream processes (H1 &
H2)
Instream leaf stocks decreased with FP and increased with %UC. The
negative effect of FP on instream leaf stock was stronger in autumn than
in other campaigns (Table 2, Fig. 2a). Leaf C/N content increased (i.e.
quality decreased) with FP (Fig 2b), but mainly so during spring and
autumn, while it decreased with DS in spring and summer (Table 2, Fig
2c). %N showed opposite patterns to C/N and likely drove changes in
C/N, whereas %C increased with FP and %UC (Table S3). Leaf C/N was
lower in summer and autumn than in spring, while leaf stock was higher
in autumn than in other seasons (Table 2).
Instream invertebrate richness and leaf-shredder abundance increased
with FP (Fig. 2d) and this effect was more pronounced as DS increased
particularly for richness (Fig2e) but less so for shredder abundances
(Table 2). Shredder abundance also increased with %UC in spring and
autumn (Fig 2f), whereas richness decreased with increasing C/N, i.e. as
leaf quality decreased (Table 2).
CK increased with FP (Fig. S5) and this effect was mediated through
changes in shredder abundances, when those were included in the models
(Table 3). CK also increased with UC% and data exploration revealed
that an orthogonal quadratic (Table S4) (i.e. bell shaped) regression
(AICc = -510.9) was a better fit than linear (AICc = -503.5),
particularly at low DS (i.e. among headwaters; Table S4, Fig 2g). An
even better set of models (AICc = - 518.0), including network location
(categorical) as interacting factor with DS (continuous), revealed that
CK increased with DS among headwaters only (Fig 2h; Table S4).
Decomposition rates in fine mesh bags (FK) increased with DS in summer
(Table 3, Fig. 2i) and slightly decreased with leaf-stock quantity.