One-dimensional stepping stone models, sardine genetics and Brownian local time (Q2476408)

A one-dimensional stepping stone model is considered in which \(L\) colonies with \(M\) haploid individuals are placed equidistantly on a straight line. Nearest neighbor migration occurs with the probability \(\nu\) per generation. The asymptotic distribution of the coalescence time \(t_0\) of two lineages, one sampled from the colony at 0 and another from the colony at \(L-1\), is investigated. It is shown that if \(L\to\infty\) and \(M\nu/L\to\alpha\in(0,+\infty)\), then \(2t_0/(L^2/\nu)\) converges in distribution to \(l_0^{-1}(\alpha\xi)\), where \(l_0\) is the local time at 0 for a standard Brownian motion starting at 1 and \(\xi\) is independent with a mean 1 exponential distribution. An analogical result is obtained for a sequence of voter models on \(Z\). A parabolic differential equation is described for computing the limiting distribution.