Spatiotemporal response of the male Kirtland's warbler population to changing landscape structure over 26 years

The interaction between population dynamics and landscape change over an extended period of time and large geographical area has important implications for habitat management and conservation, but is rarely considered in many studies. I evaluate how changing landscape structure from forest management and wildfire disturbances influenced population size, growth and spatial distribution within their restricted breeding range in northern Lower Michigan. Kirtland's Warblers breed exclusively in early-successional jack pine forests. From 1979 to 2004, the male population increased five-fold (205--1322), while suitable habitat more than doubled on the landscape. At the regional scale, significant points of change in the population trend were identified in 1987 and 1994. In 1987, the population trend began increasing from a slowly declining trend in 1987, while the rate of increase appeared to slow after 1994. The overall amount of suitable habitat and the relative area of wildfire-regenerated habitat were the most important factors explaining the population trend. Spatiotemporal distribution of males shifted across the landscape in response to the amount and location of wildfire habitat. At the patch-scale, the male population responded differently to patch area and isolation across periods of varying male population and proportion of suitable habitat. Probability of patch occupancy significantly increased with patch area and significantly decreased with isolation, but the relationship was non-linear in some periods. Male density at the patch-scale declined linearly with isolation. Examination of the timing of patch colonization and abandonment found larger patches were colonized typically earlier and abandoned later than smaller patches, and isolated patches were generally colonized at a later age. Patch colonization was earlier as the fraction of occupied patches increased on the landscape, while colonization occurred later during periods of rapid habitat influx. Isolated patches were abandoned at earlier ages. Patches were abandoned at older ages as the fraction of occupied patches increased, and abandoned at younger ages during rapid habitat influx. The long-term temporal scope and large-scale spatial scale of this study furthers our understanding of the consequences of patch dynamics on migrant bird populations within the context of a habitat restoration program.