Fig. 4

a), b) and c) Median dates of the pre-nuptial migration. Dates indicate when 50% more encounters than expected from the capture/observation of stationary individuals do occur; they have been calculated on the whole (a) or the rarefied (b) ringing dataset, or the eBird dataset (c). Isolines represent areas where the migration date occurs at the same time (e.g. in (a) in the central-southern part of Sicily the median date of pre-nuptial migration occurs during the first ten-day period of February, while in the rest of the island, it occurs mainly during the second ten days of the same month). Months are divided into ten-day periods (‘decades’ sensu the Key Concepts Document of the EU Birds Directive; e.g. Jan 1, Jan 2, Jan 3). Isolines labels should be interpreted as the first day of the corresponding decade, e.g. isoline JAN1 should be read as “01 January”, JAN2 as “11 January”, JAN3 as “21 January” and so on. d), e) and f) Sensitivity analyses associated with panels a), b) and c), respectively. Isolines include areas with the same sensitivity values (in days) of the estimated median date of the pre-nuptial migration