Fig 1.
Location of the LTER Plaine et Val de Sèvre (study area) and positions of the 260 point count stations.
These stations are distributed fairly evenly over the whole study area.
Fig 2.
Principles and comparison of the two methods: metacommunity pattern analysis characterizes the patterns generated by environmental factors, biotic interactions or random processes whereas variation partitioning can distinguish the effect of environment factors and dispersal on the community assembly.
Fig 3.
Global variation partitioning analysis (all species classes and all landscape classes). a. In percentage of the total variation b. In percentage of the explained variation. The variation explained by each variable (geographical = blue curve, environmental = green, temporal = red curve), and interaction are represented with respect to the total explained variation (black curve), at each environmental grain (x-axis). The barplot represents the relative effect of each environmental predictors based on the F-ratio.
Fig 4.
Variation partitioning per landscape and species class (% of explained variation).
The curves represent the part of the explained variation by each variable: total (black), environmental (green), geographical (blue), temporal (red) and all interactions, at each environmental grain (x-axis). The barplot represents the relative effect of each of the environmental variables (based on the F-ratio). A. Partitioning per landscape B. Partitioning per species.