TURNHOUT C VAN, E VAN WINDEN, G TROOST, K KOFFIJBERG & F HUSTINGS (2009) Changes in the timing of visible bird migration in The Netherlands in autumn. LIMOSA 82 (2): 68-78.

Counts of visible bird migration have a long tradition in The Netherlands, both regarding movements of seabirds and movements of migratory birds over land. Counting effort of migration over land has varied throughout the years, peaking in the 1980s when studies were coordinated by the national working group LWVT and after 2000, when possibilities to share data via the internet (www.trektellen.nl) have stimulated a growing number of birdwatchers to establish observation sites and count migration (Fig. 1). Meanwhile a long term data series has been established, covering the period 1980-2006. We analysed these data in order to detect possible longterm changes in timing of migration. For this purpose migration patterns in 37 species during autumn migration (period 10 June - 1 December) were analysed. The dataset included counts of 64 observation sites that mainly covered inland broad-front migration and where 20 or more counts had been carried out annually in at least three years during 1980-2006.
      According to their median dates, Common Swift (14 July) and Bean Goose spec. (14 November), were the first and last species to pass in autumn (Fig. 3). The majority of species (21 out of 37) showed median dates in the first three weeks of October. Overall, a slight advancement in migration patterns of 0.11 days per year was found, i.e. three days in the period 1980-2006. It was mainly the start of migration that advanced (Fig. 4a) whereas the end of migration activity was delayed (Fig 4b), extending the main migration period. There were no consistent differences between long-distance (Trans-Sahara) migrants and migrants moving over shorter distances (Southwestor Northwest-Europe). Median dates advanced in 21 species, though (nearly) significantly (P<0.1) only in five species. Median dates retarded in 16 species, of which (nearly) significantly in five. (Table 1, Fig. 4).
      These findings partly contradict earlier studies from abroad, but this may partly be explained by different sets of species and years under study. This could be resolved by international collaboration and coordination. More problematic is the long term shift in counting methods in The Netherlands, from very standardised to mainly unstandardised. For instance, recent tendencies to concentrate counting effort in the most profitable periods (Fig. 2) may obscure trends in timing of migration. Therefore, we strongly recommend reintroducing some standardi - sation in the current counting methods in order to allow proper comparisons and enhance data analysis.

[pdf only for members] [dutch summary]