POPULATION CONSEQUENCES OF WINTER HABITAT LOSS IN A MIGRATORY SHOREBIRD .2. MODEL PREDICTIONS

1. This paper models empirically how habitat loss in winter might affe ct the size of the European population of oystercatchers Haematopus os tralegus ostralegus. It explores how a density-dependent mortality rat e in winter interacts with a density-dependent production rate in summ er to determine the total, year-round population size following a loss of winter habitat which itself leads to intensified competition for f ood and hence increased winter mortality rates. 2. Simulations over a range of probable parameter values show that the density at which wint er mortality becomes density-dependent, cW, simply determines the poin t at which population size is affected as habitat is gradually removed . The population is affected sooner in the more widely fluctuating Con tinental subpopulations than in the less variable Atlantic subpopulati ons. 3. Once winter density reaches cW, the consequences depend on the slope, bW, of the density-dependent winter mortality function. In all subpopulations, the reduction in population size increases sharply as bW increases, but only at low values; above a certain level, further increases in bW make less difference. Because of their higher reproduc tive rate, inland subpopulations are initially less affected by winter habitat loss than coastal subpopulations. These conclusions are robus t over a range of assumptions about competition for territories in sum mer and age difference in mortality in winter. 4. Adding density-depen dent hedging success to the basic model reduces the effect of winter h abitat loss on population size, but only when low proportions of the h abitat are removed. A higher mortality rate in females, whether only i n postfledging young birds or in birds of all ages, makes little addit ional difference to the population consequences of habitat loss. 5. Fi eld studies on winter habitat loss in migratory bird populations shoul d first test whether density has already reached the critical level, c W; i.e. whether some birds already die of food competition. The parame ter bW should then be estimated to determine whether its probable valu e lies in the range within which predictions are sensitive or insensit ive to its precise value. Whether the summer density-dependent functio ns are linear or curvilinear needs also to be explored, as does the ef fect of interactions between subpopulations which have different fledg ling production rates but share the same winter habitat.