Nearly the entire (98%) and over a third of the East Asian population of Brent Geese stage in Notsuke Bay and nearby sites of eastern Hokkaido, Japan, during autumn and spring migration, respectively,. Despite the region’s importance as a refueling site for migrating Brent Geese, little is known about how these migrants use specific sites within this region. In this study, we investigated the seasonal variation in the movements and use of four key sites in the eastern Hokkaido using radio telemetry. Notsuke Bay was the primary staging area for Brent Goose in both seasons, but there were frequent movements between Furen Lake in fall and between Kunashiri Island in spring. The three sites lie in close proximity to one another in the Nemuro Strait, enabling relatively quick (<30 minutes) flights between sites. Consequently, these sites may be considered as one continuous habitat during migration. Brent Geese primarily foraged on eelgrass beds at night in fall and during the day in the spring. The occupancy rates of marked geese within the protected areas of Notsuke Bay and Furen Lake were 74% and 71% during autumn, and 64% and 87% during spring, respectively. However, extending the buffer to 3 km from the boundaries of the protected areas resulted in residency rates exceeding 90% during both seasons. To effectively conserve these important staging sites in the Nemuro Strait, we recommend expanding the protected areas and including connecting waters between Notsuke Bay and Kunashiri Island and Furen Lake.

1. Researchers generally ascribe demographic drivers in a single or few sub-populations and presume they are representative. With this information, practitioners implement blanket conservation measures across metapopulations to reverse declines. However, such approaches may not be appropriate in circumstances where sub-populations are spatiotemporally segregated and exposed to different environmental variation. 2. The Greenland White-fronted Goose Anser albifrons flavirostris is an Arctic-nesting migrant that largely comprises two sub-populations (delineated by northerly and southerly breeding areas in west Greenland). The metapopulation has declined since 1999 but this trend is only mirrored in one sub-population and the causes of this disparity are unclear. Here we compare the drivers and trends of productivity in both sub-populations using population- and individual-level analysis. 3. We examined how temperature and precipitation influenced population-level reproductive success and whether there was a change in the relationship when metapopulation decline commenced. In addition we used biologging devices to reconstruct incubation events and modelled how phenology and environmental conditions influenced individual-level nest survival. 4. Correlations between reproductive success and temperature/precipitation on the breeding grounds have weakened for both sub-populations. This has resulted in lower reproductive success for the northerly, but not southerly breeding sub-population, which at the individual-level appears to be driven by lower in nest survival. Earlier breeding ground arrival and less precipitation during incubation increased nest survival in the northerly breeding population, while no factors examined were important for the southerly breeding sub-population. This suggests reproductive success is now driven by different factor(s) in the two sub-populations. 5. Demographic rates and their environmental drivers differ between the sub-populations examined here and consequently we encourage further decomposition of demography within metapopulations. This is important for conservation practitioners to consider as bespoke conservation strategies, targeting different limiting factors, may be required for different sub-population.