I developed a demographic model for species inhabiting a mosaic of habitats
of varying quality and among which dispersal was limited only by territori
al behavior I used the model to describe an ongoing decline of an Eastern K
ingbird (Tyrannus tyrannus) population breeding in the Charlotte Valley of
central New York and to explore the possibility that the population utilize
d a network of source-sink habitats Population size declined in upland and
floodplain habitats between 1989 and 1999, but remained stable along a ripa
rian zone (the "creek"). Population size after 10 years was predicted accur
ately when I used the empirically observed habitat-specific rates of adult
survival, productivity and adult dispersal among habitats and assumed a 5%
rate of immigration into the upland. Simulation of the effect of density-de
pendent variation in reproductive, success or a small increase in adult sur
vival in all habitats failed to improve the model's predictive ability. The
model indicated that none of the subpopulations teas self-sustaining over
50 years and that population decline was driven by declining adult survival
. Further analyses, however, showed that small increases in adult survival
along the creek! changed the creek habitat into a source and reversed popul
ation trends. Simulation of realistic rates of habitat loss as a result of
forest maturation did not affect predicted population size at 50 years, but
, assuming that the area of pasture and grazed lands continues to decline a
t current rates, the king bird population is predicted to become extinct lo
cally within 40 years. Riparian zones appear to be source habitats for king
birds and should be targeted for protection. Sink habitats were also import
ant because in every year most of the king bird population bred on the floo
dplain or in the upland and many floodplain birds later dispersed to the cr
eek. Preservation of open space is thus critical for maintenance of kingbir
d populations in eastern North America.