Meta-analysis
Kendall’s rank test for funnel plot asymmetry indicated that no
publication bias existed for all the tests (P >
0.05) (Table S1.8; Fig. S1.11), and there was no significant correlation
between effect size (Z ) and studies’ publication years/journal
impact factors (P > 0.05) (Fig. S1.12).
Overall, elevation was not
associated with foliar fungal pathogen OTU richness (Z ± 95% CI
= -0.157 ± 0.263, P = 0.242), foliar fungal diseases (Z ±
95% CI = -0.047 ± 0.242, P = 0.703) or sfpRA (Z ±
95% CI =-0.101 ± 0.183, P = 0.281), whereas increasing elevation
was significantly associated with sfpOTUs (Z ± 95% CI =
-0.257 ± 0.172, P = 0.003) (Fig. 3; Fig. S1.13; Table S1.9).
Elevation was not significantly
associated with foliar fungal disease in forest (Z ± 95% CI =
-0.127 ± 0.440, P = 0.571) or grassland (Z ± 95% CI =
0.023 ± 0.153, P = 0.772) ecosystems (Fig. 3; Fig. S1.13; Table
S1.9). There was no significant correlation between the climatic
variables (mean annual temperature, mean annual precipitation) and the
effect size (Z ) on ffpOTUs , foliar fungal diseases,sfpOTUs or sfpRA (Fig. S1.14; Table S1.10). Nevertheless,
increasing absolute latitude was associated with decreasing effect size
(Z ) on ffpOTUs (Q m = 4.231,P = 0.040) but not foliar fungal disease
(Q m = 0.364, P = 0.547), sfpOTUs(Q m = 1.264, P = 0.261) or sfpRA(Q m = 1.647, P = 0.199) (Fig. S1.14; Table
S1.10). Furthermore, studies
conducted over larger elevational ranges tended to observe stronger
negative relationships between elevation and sfpOTUs(Q m = 4.572, P = 0.033) and sfpRA(Q m = 4.469, P = 0.033), though the
elevational range of sampling did not moderate this relationship forffpOTUs (Q m = 0.150, P = 0.698) or
foliar fungal disease (Q m = 0.415, P =
0.520) (Fig. S1.14; Table S1.10).
Discussion