Studying movements and population turnover is a prerequisite of management and conservation policies. Understanding the relative importance of known wintering sites requires estimating total number of birds using a wintering site with robust statistical methods because counts alone do not account for turnover. To study movements of common pochards (Aythya ferina) during winter (Oct to Mar) in western France, we used a combination of band-recoveries and capture-recapture data. We constructed a multi-state model to estimate monthly local survival and probability of movements between the banding site and other wintering sites, accounting for individual sex and age. We observed significant movements between sites and high population turnover even during winter. We did not detect any effect of age or sex on movement probabilities, except at the beginning of the winter when first-year individuals had a lower emigration probability. This result suggests that these inexperienced birds did not explore the wintering area like adults, at least early in the season. Combining our estimates (survival and emigration) and winter counts data, we computed the number of individuals using our study area throughout winter. This total was twice the maximum number of individuals counted in winter. Unexpectedly, low estimates for apparent survival suggest that permanent emigration occurred; although, recaptures and resightings revealed some birds traveled long distances from the capture site. We hypothesize that this permanent emigration was due to birds moving to areas where observation and hunting pressures were reduced compared to those in the French territory. If this hypothesis is true, high turnover rate would prevail not only at the scale of the wetland but also at the scale of the country. \textcopyright The Wildlife Society, 2013.