Figure 4
Dispersal kernel instead of dispersal threshold. (a) nodes with their respective coordinates linked by a probabilistic dispersal network. Link width is proportional to the frequency at which two nodes are connected when drawing random dispersal events following a Weibull distribution (methods). Node colors represent the module partition that maximize modularity in weighted networks. (b) Posterior dispersal kernel obtained from panel a. It shows the dispersal probability as a function of distance between nodes. The arrow points the maximum dispersal distance recorded (c) Convergence of the value of modularity as a function of the number of observations. Each observation is a binary network where a link is established as a function of a probability dispersal kernel. The number of observations represent how many of these networks we add together to create the weighted network over which we calculate modularity. The distributions correspond to combinations of 100 weighted networks. (d) Analogous to Fig. 2a. It shows how modularity decreases as a function of mean of the Weibull distribution used to determine whether a link is established for 100 random landscapes. For comparison purposes, each of the individual binary networks have the same connectance that those in Fig. 2a and the mean dispersal distances coincides with the threshold distance.
