Seasonal mortality and sequential density dependence in a migratory bird

Migratory bird populations may be limited during one or more seasons, and thus at one or more places, but there is a dearth of empirical examples of this possibility. We analyse seasonal survival in a migratory shellfish-eating shorebird (red knot Calidris canutus islandica) during a series of years of intense food limitation on the nonbreeding grounds (due to overfishing of shellfish stocks), followed by a relaxation period when destructive harvesting had stopped and food stocks for red knots recovered. For the estimation of seasonal survival from the 15 yr-long near-continuous capture-resight dataset, we introduce a 'rolling window' approach for data exploration, followed by selection of the best season definition. The average annual apparent survival over all the years was 0.81 yr(-1). During the limitation period, survival probability of adult red knots was low in winter (0.78 yr(-1)), but this was compensated by high survival in summer (0.91 yr(-1)). During the relaxation period survival rate levelled out with a winter value of 0.81 yr(-1) and a summer survival of 0.82 yr(-1). The fact that during the cockle-dredging period the dip in survival in winter was completely compensated by higher survival later in the annual cycle suggests sequential density dependence. We conclude that seasonal compensation in local survival (in concert with movements to areas apparently below carrying capacity) allowed the islandica population as a whole to cope, in 1998-2003, with the loss of half of the suitable feeding habitat in part of the nonbreeding range, the western Dutch Wadden Sea. As a more general point, we see no reason why inter-seasonal density dependence should not be ubiquitous in wildlife populations, though its limits and magnitude will depend on the specific ecological contexts. We elaborate the possibility that with time, and in stable environments, seasonal mortality evolves so that differences in mortality rates between seasons would become erased.