Influence of coloured noise on the extinction risk in structured population models

We use simple models to compare extinction risk among basic life history types when environmental noise is either uncorrelated (“white”) or positively autocorrelated (“red”). The metric of extinction is probability of extinction in 50 years; variability of noise is scaled such that its expected variance is independent of colour at this time scale. We compare annual, semelparous biennial, iteroparous biennial and perennial life histories. Given an identical equilibrium population size and basic reproductive number, annual life history confers much higher extinction risk than semelparous biennial life history. Iteroparous biennial and perennial life histories have even lower extinction risks. Autocorrelated noise influences the life histories differently: the extinction risk of the annuals decreases with reddening noise, whereas the other life histories show an opposite response. We show that in a previously developed stage-structured population model for the Florida scrub jay Aphelocoma coerulescens ignoring the possibility of red environmental noise might result in much too optimistic assessment on population viability. In conclusion, simplifications of population structure and ignoring red environmental noise in population viability analyses can result in serious biases.