Stopover departure decisions in spring: pre-Saharan migrants stay longer and are more selective for favourable wind than trans-Saharan migrants

Abstract

Background: Birds that breed in Europe and winter south of the Sahara, so-called trans-Saharan migrants, generally migrate longer distances than pre-Saharan migrants. The latter are expected to be less time constrained during autumn migration than the former. As such, pre-Saharan migrants are assumed to be more selective for favourable weather conditions and are more likely to minimise energy cost of migration than trans-Saharan migrants. While this pattern is supported for autumn migration, it is less well understood for spring migration. Since the optimal arrival timing at the breeding areas is generally under selection pressure to arrive ‘early’, i.e. before ‘competitors’, and since this advantage is likely to hold across migration strategies, we predict that the general differences in decision making between pre- and trans-Saharan migrants will also be manifested during spring migration.

Methods: We radio-tracked three trans-Saharan (Common Redstart Phoenicurus phoenicurus, Garden Warbler Sylvia borin and Northern Wheatear Oenanthe oenanthe) and two pre-Saharan (European Robin Erithacus rubecula and Dunnock Prunella modularis) migrants during stopover using a regional network of Motus receiving stations. We analysed the night-to-night and within-night departure decisions in relation to weather and energy stores, and compared species’ departure direction with the location of their ring recoveries.

Results: Trans-Saharan migrants stopped-over shorter and were less selective for favourable wind conditions than pre-Saharan migrants. The positive effect of high energy stores and low cloud cover on departure probability was a consistent pattern. Within-night departure times did not differ between migration strategies. Departure directions were in line with geographical mean location of ring recoveries for Common Redstart, European Robin and Dunnock.

Conclusions: Our results suggest that pre-Saharan migrants are less time-constrained and follow an energy-saving strategy more strongly than trans-Saharan migrants that seem to have a stronger urge to migrate fast in spring. Since a similar pattern exists for autumn migration, we suggest that how the species-specific migration strategies and associated time constraints affect stopover decision making in both migration seasons is a general mechanism in migratory songbirds.

Supplementary Information: The online version contains supplementary material available at 10.1186/s40462-025-00575-0.

Keywords: Departure decision; Departure direction; Energy stores; Radio-tracking; Songbird; Weather conditions; pre-Saharan migrants; trans-Saharan migrants.

Fig. 1

Distribution maps and ring recoveries of the study species. Portraits of species are given above their corresponding, restricted distribution maps (provided by Birdlife [51]). ‘All year’-occurrence (green) does not necessarily mean that the European Robin and Dunnock are resident in these areas but that some breeding birds migrate south-west in the autumn and their ‘breeding areas’ are filled by conspecifics from more northerly populations during winter. Black arrows indicate ring recoveries from birds caught on Helgoland (red dot) during spring migration (March-May) and were re-sighted in the same year until the end of the breeding season (end June). The red arrow indicates the geographical mean location of the ring recoveries, see Fig. S1 for more details. Bird photos: TKa.

Fig. 2

Prediction (i): Variation in minimum stopover duration. The three trans-Saharan songbird species (above dashed line from top to bottom: Redstart, Garden Warbler, Wheatear) show significantly shorter minimum stopover duration than the pre-Saharan migrants (below dashed line: Robin, Dunnock) during spring migration. Box plots show the 5th, 25th, 50th, 75th and 95th percentile as well as outliers (dots). Bird photos: TKa

Fig. 3

Prediction (ii): Night-to-night departure probability predicted by time-dependent Cox proportional hazards models. Predictions (lines) and associated 95% confidence intervals (shaded areas) are given. (A, C-G) Colour scheme corresponds to species: Redstart (yellow-orange), Garden Warbler (light blue), Wheatear (grey), Robin (red) and Dunnock (brown). (A) Species-specific differences in the night-to-night departure probability. (B) All species were combined to visualise the effect of cloud cover on the night-to-night departure probability. Minimum (broken line; clear sky) and maximum (solid line; overcast) of the scaled cloud cover birds experienced at sunset were used. (C-G) Species-specific effect of wind condition on night-to-night departure probability, given for the 25th percentile (broken line; headwind) and 75th percentile (solid line; tailwind) of the scaled air speed flow assistance birds experienced at sunset. A plot depicting the effect of Julian day is provided in Fig. S4. Bird photos: TKa

Fig. 4

Prediction (iii): Variation in nocturnal departure timing as observed in five songbird species during spring migration. Box plots show the 5th, 25th, 50th, 75th and 95th percentile as well as outliers (dots). Species displayed are, from left to right: Redstart, Garden Warbler, Wheatear, Robin and Dunnock. (A) Nocturnal departure timing is expressed as proportion of night at departure. (B) Nocturnal departure timing is expressed as sun´s elevation at departure. A figure depicting the nocturnal departure time in minutes after sunset can be found in Fig. S8. Bird photos: TKa

Fig. 5

Prediction (iv)+(v): Energy stores influencing songbirds´ departures. Effect of fuel load on (A) departure decision and (B) on departure time within the night of birds departing in the first night after radio tagging. Data are aggregated over the species Redstarts, Garden Warblers, Wheatears and Robins. Dunnocks were excluded as only one bird departed in the first night. Bird photos: TKa

Fig. 6

Prediction (vi): Directional decision from the stopover site. Migratory direction from Helgoland as derived from radio telemetry (filled coloured circles, coloured arrows and coloured dashed confidence intervals) and ring recoveries (open circles, black arrows and black dashed confidence intervals) from five songbird species: (A) Redstart, (B) Garden Warbler, (C) Wheatear, (D) Robin and (E) Dunnock. Arrows and their confidence intervals (dashed lines) are only shown for significantly oriented datasets. Correspondingly coloured numbers indicate sample size. Pointy triangle indicates geographical mean of ring recoveries, see Fig. 1 or Fig. S1 for a map. Bird photos: TKa

Fig. 7

Prediction (vii): Directional consistency of migratory flights. Consistency was analysed by comparing the departure direction from Helgoland within the first ∼ 1–10 km with the bearing of the passage location on the coast after 50–100 km for five migrant songbirds, Redstart (orange circle), Garden Warbler (blue square), Wheatear (grey pointed downward triangle), Robin (red rhombus), and Dunnock (brown pointed upward triangle). The black diagonal line marks the 1:1 ratio, meaning that the bird did not change its flight direction. Points are slightly (< 5°) jittered to prevent overlap. For geographical map of the German bight see Fig. S2. Bird photos: TKa