Landscape ecology and metapopulation ecology share a common interest in dev
eloping measures that describe the structure of heterogeneous landscapes, b
ut the specific aim in metapopulation ecology is to construct measures that
help predict the dynamics of species with information about the structure
of fragmented landscapes. The amount of habitat that individuals in a metap
opulation have access to can be divided into four components, the amount of
habitat in the present habitat parch (A(i)), the amount of connected habit
at in other patches available via migration (Gamma(i)), the amount of prese
nted habitat in the present patch after time period Delta t(A(i)'). and the
amount of linked habitat in other patches after time period Delta t(Gamma(
i)'). Deterministic threshold conditions for metapopulation persistence in
patch networks can be approximated with these quantities. For instance, in
a version of the Levins model with er;tinction risk proportional to the inv
erse of patch area and colonization probability proportional to patch conne
ctivity, the threshold condition for metapopulation persistence is given by
<(Gamma A)over bar> +Var(Gamma A)/<(Gamma A)over bar> > e/c, where e and c
are the species-specific extinction and colonization parameters. I conject
ure that with measures A' and Gamma' the threshold condition for metapopula
tion persistence can be extended to dynamic landscapes, in which all or par
t of population turnover is caused by turnover in the habitat patches thems
elves. The measures of habitat availability described in this paper can be
used to rank dissimilar fragmented landscapes in terms of their capacity to
support a viable metapopulation.