Dispersal kernels as a function of distance.

<p><b>A–D</b> Estimated kernels for the exponential model (orange lines) and the Cauchy model (cyan lines), plotted together as a function of distance for each census site. The functional form of the kernels is based upon <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1003587#pcbi.1003587.e009" target="_blank">Equations 3</a>, with a cut-off at very short distances as explained in <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1003587#pcbi.1003587.s012" target="_blank">Text S1</a> (Equations S5). The mean of the posterior distribution for α for model (<i>ΔT</i> = 6 months; cf. <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1003587#pcbi-1003587-g003" target="_blank">Figure 3A</a> for the exponential model) is used as a point estimate to plot each kernel. The value of the two kernels at the point where they begin to diverge (“plus” symbol in <b>A–D</b>) is about 10³ times the value at very short distances.