Demographic heterogeneity has long been considered within wildlife populations, but only the modern development of capture–mark–recapture methods allows this to be easily tested and quantified. It is now possible to rapidly assess whether the modeling of heterogeneous populations, in which categories of individuals differ in survival rate, performs better than traditional approaches, in which all individuals are considered equivalent within a sex and age class. Using long-term banding data for 4,703 adult female Green-winged Teal (Anas crecca) from the Camargue, southern France, we show that a heterogeneous model outperformed a homogeneous model. Individuals from the high survival category had a ∼60% annual survival rate, whereas birds in the second category had a survival rate reduced by a factor of 0.76–0.80, depending on the model (i.e. <50%). We could not demonstrate that individuals within the high survival category were larger or heavier. The link between survival rate and potential differences in individual morphometrics or individual behavioral strategies thus remains to be established. Previous studies in which a Green-winged Teal population was modeled as homogeneous suggested it should decline (population growth rate <1), which we also found when using demographic parameters obtained from a homogeneous model. This finding contradicts waterfowl surveys that show a long-term population increase in this flyway. Modeling the population as heterogeneous led to growth rates of 1.03–1.05 (a 3–5% annual increase), numbers consistent with the growth rate inferred from duck counts and that also partly explain how species such as Green-winged Teal can increase in numbers despite large hunting harvest, sustaining harvest to some extent.