stochastic variance in fecundity

[joelpick]

currently assumes annual fecundity is poisson, but maybe better to have possibility of underdispersion with generalised poisson? poison unrealistic for most systems