Issue warning when a species abundance is not negligible above its w_max

[gustavdelius]

Now that we allow w_max to be less than w_repro_max, it is no longer guaranteed that the population density drops to zero beyond w_max. Mizer should issue a warning when w_max is smaller than w_repro_max and N(w_max) is not negligibly small, say less than 10^-6 * N(w_mat).