Light-level geolocators were used to record the annual migration cycles of black terns Chlidonias niger (9 inds, 11 journeys) and common terns Sterna hirundo (7 inds, 11 journeys) breeding in southernmost Sweden. The black terns used two different winter (Oct – Mar) regions along the Atlantic coast of Africa, either north of the equator between Senegal and Liberia (3 inds) or south of the equator between Gabon and northern Namibia (5 inds). All the common terns travelled to winter quarters south of the equator, mainly along the coasts of Namibia and South Africa. One juvenile common tern was tracked during the first twenty months of its life. This bird spent its first winter at South Africa, after which it migrated north of the equator to over-summer as one-year old non-breeder in tropical waters off Ghana, after which it returned to South Africa for its second winter. This record demonstrates that one-year old terns may undertake extensive intra-African migration to distant over-summering areas. Comparing geolocator results from Swedish and Dutch black tern populations indicate that they have similar migration habits, with a possible tendency of relatively more individuals migrating south of the equator in the more northerly Swedish population (leap-frog migration). Comparing geolocator and ringing results among common tern populations indicates a fascinating and complex pattern of scale-dependent geographic segregation and intermixing along the coasts of Africa.  

We used geolocators manufactured by Migrate Technology Ltd (Intigeo series; Fox 2021) and British Antarctic Survey (BAS). The attached geolocators weighed 1.5-2.0 g (including glue, string and/or plastic ring), which corresponds to ≤ 3% of the body mass of the black tern (average body mass = 65g) and < 2% of the body mass of the common tern (average body mass = 137 g).

During 2014 – 2021 a total of 69 black terns were equipped with geolocators at two breeding areas in the southernmost province of Sweden (Scania): Kristianstads Vattenrike (northeastern Scania), and the lakes Krankesjön and Börringesjön (southwest Scania), using a leg-loop harness. Out of these 15 birds were recaptured and the loggers removed. However, since 6 of the loggers were lost in the mail, this data sets includes data from 9 black terns.

66 common terns were equipped with geolocators 2009 – 2015 in two different areas of southernmost Sweden: at the coastal archipelago of northeastern Scania and at the Lake Krankesjön. 22 of the loggers were attached on a color ring on the leg of the bird, and 44 loggers were attached using a leg-loop harness. A total of 8 loggers with useful data were retrieved. Two of the loggers were from the same bird with repeated journeys (1,5 + 1 year).  One retrieved logger was from a juvenile bird that got a logger just before it fledged at Lake Krankesjön on 27 Jul 2013. It was found dead at Fisherhaven (34.355⁰S, 19.118⁰E) in the western Cape Province of South Africa on 25 Jan 2018. Useful data were obtained from the first 20 months (1,5 year) of the juvenile bird’s life. Hence, in total we data from six adult common terns and one juvenile is included in the data sets.

We used the IntiProc software (Migrate Technology Ltd; Fox 2021) including the Geolight package (Lisovski & Hahn 2012) for converting light data into geographical positions. With the software calibration function we used clean light data from the periods when birds were in the known breeding area for calibration of the sun angle versus light threshold for obtaining latitude. Reliable latitude data were missing from a period of about a month or more around autumn and spring equinox. These invalid latitude data were deleted based on inspection of the annual records of latitude estimates for each individual. However, longitude data were available throughout these equinox periods and could be used for some but not all timing estimates of events that occurred during the equinox periods. Daily light and location data were inspected for each individual, and data from days/periods with uncertain or invalid assignments of light data to sunset/sunrise were omitted.