 
Summary: Estimating the variance of survival rates and fecundities
INTRODUCTION
Estimating the risk of extinction or decline of a popula
tion in a population viability analysis often requires esti
mates of the variability in vital rates, such as survival
and fecundity. A common type of data used in such esti
mates consists of regular censuses of a population at
which individuals in different age classes or stages are
counted. Such data are used to calculate survival rates
and fecundities. Repeated estimates of these rates in time
allow estimating the temporal variance of these rates.
However, this observed variance includes variance due
to demographic stochasticity and measurement error.
These components must be removed from the observed
variance to estimate the variance due to environmental
stochasticity. Otherwise, the variances, and hence risks
of decline and extinction, may be overestimated.
Several previous studies have focused on the related
problem of removing sampling variance from estimates
of abundance (e.g. Link & Nichols, 1994; McArdle &
