Oikos 118: 624Á626, 2009 doi: 10.1111/j.1600-0706.2009.17309.x, # 2009 The Authors. Journal compilation # 2009 Oikos Subject Editor: Hamish McCallum. Accepted 3 November 2008 The natural link between Europe and Africa 2.1 billion birds on migration Á Steffen Hahn, Silke Bauer and Felix Liechti S. Hahn ([email protected]), S. Bauer and F. Liechti, Swiss Ornithological Inst., Luzernerstrasse 6, CHÁ6203 Sempach, Switzerland. SB also at: Dept of PlantÁAnimalÁInteractions, Centre for Limnology, Netherlands Inst. of Ecology (NIOO-KNAW), PO Box 1299, NLÁ3600 BG Maarssen, the Netherlands. The PalaearcticÁAfrican migration system comprises enormous numbers of birds travelling between Europe and Africa twice each year. Migratory birds may form strong links between the two continents given they can act as both transport vehicles for parasites and diseases as well as temporary consumers with increased food demand to fuel their flight. Knowing the number of migrating birds is crucial if such links are to be quantified. We estimate that today approximately 2.1 billion songbirds and near-passerine birds migrate from Europe to Africa in autumn, 73% of which are accounted for by just 16 species. This number is only half the estimate from the 1950s in the only other assessment to date. The discrepancy is mainly caused by the limited information on population sizes in the past. Our estimated number of migrants is highly dependent on the accuracy of the underlying estimates of breeding population sizes, as well as breeding parameters in species with relatively high reproductive output. The updated figures quantify and emphasize the strong natural connection between Africa and Europe, which has important implications for manifold research topics including those related to climate change, human health and biological conservation. ‘‘Let us now toss a few figures into the air.’’ With these between Europe and Africa in boreal autumn. Additionally, words Moreau (1972, p. 45) introduced his attempt to we reconsidered Moreau’s calculations to identify the cause estimate the number of birds involved in the world’s biggest of potential differences Á be they methodological or due to bird migration system, the PalaearcticÁAfrican flyway actual changes in populations sizes. (Newton 2008). Migratory birds connect the two con- tinents and their different biomes twice each year. Through the transportation of disease (Hubalek 2004) and disease Methods vectors (Comstedt et al. 2006), parasites (Waldenstrom et al. 2002), seeds (Sanchez et al. 2006), and the harvesting We focused on the autumn migration of birds from their of considerable amounts of food while preparing for flight European breeding grounds, within the geographical (Alerstam and Lindstro¨m 1990, Fransson et al. 2008) they borders of Europe including Turkey and European Russia seasonally alter the composition of local communities at (to approx. 608E), to their sub-Saharan African wintering their sites of origin, as well as stop-over and destination quarters. Bird populations were assumed to include sites. Estimating the number of birds migrating between breeders, their offspring and non-breeding adults. We Europe and Africa is of utmost importance when we strive considered 55 passerine and 13 near-passerine species, i.e. to quantify such links. Eurasian wryneck Jynx torquilla, European roller Coracias However, to date, the only estimate of how many birds garrulus, Eurasian hoopoe Upupa epops, bee-eaters Merops migrate between the two continents stems from Moreau’s sp., swifts Apus sp., Tachymarptis melba, nightjars Capri- pioneering work (Moreau 1972) that deduced the impress- mulgus sp., common cuckoo Cuculus canorus, European ive number of 5 billion migrating landbirds, dominated turtle dove Streptopelia turtur and common quail Coturnix by 4.3 billion passerines. Although this figure has been coturnix, all migrating to sub-Saharan regions (Supple- heavily used in subsequent works, it has not been revisited mentary material Appendix 1, Table A1). We used since its publication more than 30 years ago. published estimates of current population sizes of birds We used current data on breeding population sizes, breeding within the above mentioned region (BirdLife reproductive success, proportions of young and non- International 2004). We determined the proportion of breeders of species from Europe including Turkey and long-distance migrants within each population using European Russia to establish a current estimate of the information derived from Glutz von Blotzheim and Bauer number of passerine and near-passerine birds migrating (1987) and Cramp (1998). In five species, bluethroat 624 Luscinia svecica, common chiffchaff Phylloscopus collybita, Results Eurasian blackcap Sylvia atricapilla, Eurasian hoopoe and common quail, considerable parts of the European breed- Based on the size of breeding populations of European ing population undertake partial migrations with several species (BirdLife International 2004), we estimated that in sub-populations wintering north of the Sahara. For these total, between 1.52 and 2.91 (average 2.1) billion passerine species we assumed complete residency in Mediterranean and near-passerine birds cross the Sahara each autumn to populations and/or a mixed migration strategy, with 50% spend the non-breeding season in sub-Saharan Africa. of individuals in populations along a corridor throughout Passerines account for 95% of all individuals. Moreover, central Europe over-winter north of the Sahara. The 73% of these long-distance migrants belong to 16 species resulting numbers of birds, which don’t migrate to (Fig. 1) of which the willow warbler Phylloscopus trochilus is sub-Saharan Africa, have been subtracted from their the most numerous (15.8%), followed by tree pipit Anthus European migrant population size (Supplementary Mate- trivialis (7.3%), common chiffchaff and barn swallow rial Appendix 1, Table A2). Hirundo rustica (both 6.5%). Population composition The number of first-year birds on migration was derived varied between species according to their mean breeding from data of mean annual breeding success (no. of success. On average, 47% of migrants were adults (minÁ fledglings per nest), the average number of breeding max: 30Á63%), 46% juveniles (27Á66%) and 7% non- attempts per season and the estimated survival from breeders (4Á10%, Fig. 1). fledging until departure. Data on reproduction were derived The accuracy of these estimates is principally reliant from Glutz von Blotzheim and Bauer (1987) and Cramp upon published data of the species’ breeding population (1998); some missing values were substituted with averages sizes; varying this parameter for all species by 10% would, from the closest relatives. Survival estimates for the post- of course, also change the estimated total number of fledging period were lacking for most species. Therefore, we migrants by 10%. Similarly, variations in clutch size, used a median survival rate of 0.6 for all species by taking breeding success and fledging survival entail changes of the average fledgling survival of the six passerine species for 4.9% each (at species level: minimum 2.71%, maximum: which data were available (Naef-Daenzer et al. 2001, 6.58%). Less important were the proportions of non- Kershner et al. 2004, Gru¨ebler 2007, Wells et al. 2007). breeders and resident birds in the total population, which Finally, the proportion of non-breeders within a population change the overall estimate by only 0.9% and 0.6%, was estimated to be 15%, the median of seven passerines respectively (species: minÁmax: 0.45Á0.95%). Finally, (Newton 1998). assuming that a long-distance migration strategy is em- Because parameter estimates might be subject to in- ployed by the entire population for all species would accuracy, we determined the sensitivity of the output increase the total number of migrants by 7.9%. (number of migrating birds) to variation in each parameter. Therefore, we subsequently changed all parameters by 910%, re-calculated the output and, finally related the Discussion changes in the focal parameter to the resulting changes in the output value (Hamby 1994). Moreau (1972) based his calculations on population-size All population-relevant averages are geometric means. estimates of Finnish birds from the 1950s (Merikallio willow warbler tree pipit common chiffchaff* barn swallow spotted flycatcher garden warbler common whitethroat common house martin European pied flycatcher common redstart wood warbler Eurasian blackcap* northern wheatear western yellow wagtail breeding adults greater short-toed lark juveniles common nightingale non-breeding adults 0 100 200 300 400 No. birds (× 106) Figure 1. The 16 most abundant long-distance migratory species crossing the Mediterranean Sea and the Sahara desert, representing 73% of all European passerines and near-passerines during autumn migration. Bar length gives the geometric mean number of breeding adults, their juveniles and the non-breeding adults per species; the error bars indicate minimum and maximum numbers. Asterisks mark species with partially migratory populations. 625 1958), and estimated that 4.3 billion passerines and 200 E. (ed.), Bird migration: physiology and ecophysiology. million near-passerines migrate to Africa in autumn. Our Springer, pp. 331Á351. calculations result in numbers that are only half these earlier BirdLife International 2004. Birds in Europe: population esti- estimates. For almost all of Moreau’s 16 most abundant mates, trends and conservation