Added day coverage after 1995 causes most of the apparent lengthening of spring and autumn migration passage at Hanko Bird Observatory, Finland, 1979–2022

Question: do percentile passage dates from the public daily counts of Hanko Bird Observatory (Finland, 1979–2022, FinBIF/GBIF) show longer migration seasons when the day coverage is the same in all years? The station counted on 181 days per year in 1979–1994 and on 333 days per year in 1995–2022. Most added days are in the tails of the seasons. Approach: for 76 spring and 89 autumn migrant species we computed the 5%, 50% and 95% passage dates with all observed days (naive) and with mask matching, which gives each year after 1994 the observed days of each early year in turn. We used a bootstrap over species and years. Results: the naive 5–95% season length increased by 3.63 days per decade in spring (95% CI 2.29 to 4.98) and by 2.57 in autumn (1.28 to 3.95). With matched coverage the trends were +0.85 (−0.33 to 2.02) and −0.12 (−1.24 to 0.86). Thus the coverage change causes 78% of the naive spring lengthening and all of the autumn lengthening. The spring median date advanced by 0.97 days per decade (0.23 to 1.68) with matched coverage. Long-distance migrants advanced more than short-distance migrants (−1.51 versus −0.68 days per decade). An April–May temperature model fitted to 1979–2000 predicted 36% of the observed advance to 2001–2022 for all migrants, 72% for short-distance migrants and 10% for long-distance migrants. In a planted test, the naive estimator added 5–6 days per decade to the season-length trend; the mask-matched error was below 1 day per decade. Limitations: one site; the open data combine local and migrating birds; mask matching can under-estimate shifts into days that the early years did not observe, so a true spring lengthening of about 1–2 days per decade is not excluded; it assumes that early observers did not choose days by bird numbers.

bird migration · phenology · bird observatory · sampling coverage · observation effort · migration duration · long-distance migrants · Hanko · Finland · climate change · FinBIF · percentile passage dates