Figure legends

Fig. 1. Mean detection probabilities ± 1 standard error of known long-term resident birds in each year using distinct methods (red = manually, green = MARK, blue = manually from radio-tagged birds). Sample sizes during each year are shown on the top. In year2 “(n =3)” represents the sample size of radio-tagged birds. Dashed line shows the overall mean detection (0.33). There were no clear differences in detection probabilities across years or methods.
Fig. 2. Distances moved between year i and year i +1 by individuals that moved above the average (> 300m; A) and below the average (< 300m; B). Sample sizes are shown on each map. A subset of individuals is shown with a higher definition in map C. Here, each colour represents a different individual. Individuals that do not have a line moved out of the confines of the box. Overall, individuals had a high degree of between-years site fidelity.
Fig. 3. Distance (meters) moved from year i to year i +1 according to previous age, sex, and previous residency.
Fig. 4. Change of individuals’ residency category from one year to the following year it was detected. Percentage in the square represents the proportion of individuals in each residency category at year i +1, as is observed by the width of the bands. Colours represent previous residency category.
Fig. 5. Linear correlation between the number of days an individual spent at the study site in year i and the number of days an individual spent at a location in year i +1. Colours represent residency category during year i . The dotted line represents a constant residency period during both years. Points above the dotted line represent individuals that remained longer in year i +1 than the previous year and points below the line represent individuals that remained a longer period in the previous year. R = correlation between variables and the p -value shows a significant positive trend.