Population extinctions in correlated environments

For a spatial population assemblage, extinction risk should be greatly affected by features of local population dynamics and interpatch migration patterns. In a variable environment, the magnitude of environmental correlation between local population patches map have great impact on local dynamics and thereby global extinction risk. We examined the effect of correlated environmental variation on global extinction risk in a coupled lattice model consisting of local populations governed by density dependent population growth and density independent interpatch migration. We let each local population experience a stochastic environment expressed as a variation in maximum birth rate and let this environmental variation be correlated among local populations. We simulated global population growth under different magnitudes of environmental variability, correlation of environmental variability, emigration rate and migration survival, in order to evaluate the magnitude of their effect on local population dynamics and global extinction risk. The risk of global extinction increases with increasing magnitude of environmental correlation and environmental variability. The major determinant of global extinction risk is the balance between local population variability and the synchrony in local population fluctuations. A low rate of successful interpatch migration connects the local populations to each other, exposing them to less extinction risk than when they are isolated. High levels of interpatch migration are often negative for population persistence. The reason for this is that increased migration survival causes an increased risk of population crashes, due to overcompensatory population growth. This effect is amplified by a high emigration rate. Thus, local dynamics are affected by temporal and spatial variability in birth rates as well as interpatch migration levels. An assemblage of local populations in a variable environment will suffer least risk of global extinction when environmental correlation is low and interpatch migration is moderate.