Migratory birds are capable of flying long distances across barriers without possibilities of stopping over (Gill et al. 2009). Even for small songbirds, long‐distance flights across barriers have been reported (Bairlein et al. 2012, Deluca et al. 2015). However, there are many examples of migratory bird species following alternative routes that differ from the direct path between departure and destination even though they are energetically capable of making long distance flights (Alerstam 2001). Based on ring‐recoveries and observations, all populations of red‐backed shrikes Lanius collurio from Europe are believed to initiate autumn migration with a first leg of migration to southeast Europe, and then follow the same migration corridor southwards to the Sahel (Korner‐Nievergelt et al. 2012).

According to migration theory, migration can be optimized by minimizing a simple currency such as energy cost of transport, time or mortality, or a combination, as for example total energy cost of migration (Hedenström and Alerstam 1997). If considering migratory birds to follow routes that minimize overall fuel transportation cost, birds will save energy by short flights with low fuel loads compared to long flights with large fuel loads. Therefore, birds are expected to accept a certain detour distance where shorter flight steps can be applied and avoid ecological barriers where food supply is low and more fuel is needed for successful completion of their migration (Alerstam 1979). Thus, in terms of energetics, the cost of transporting heavy fuel loads across a long barrier may be overcome by travelling along a detour with the possibility of making shorter stopovers requiring smaller fuel loads.

The adaptive value of making a detour may depend on multiple factors related to migratory travelling (avoiding hazards) as well as stopping over (associated with a suitable environment) (Alerstam 2001, 2011, Hahn et al. 2014). Factors favouring a route include potential wind assistance (Erni et al. 2005, Vansteelant et al. 2016), habitat and food availability (Alves et al. 2012, Thorup et al. 2017) whereas potential hazards include side‐wind drift, high probability of head winds, high predation risk (Klaassen et al. 2006, Ydenberg et al. 2007, Bauer et al. 2010, Gill et al. 2009), and sea and desert crossing (Biebach et al. 1986, Barboutis et al. 2011). Alternatively, detours could be constrained by historical patterns of dispersal and colonization events (Irwin and Irwin 2003).

To investigate whether detours from the direct path would be optimal in terms of energy cost of transport, Alerstam (2001) developed a theoretical framework based on flight mechanics (Pennycuick 2008) and optimality theory (Alerstam and Lindström 1990). Predictions were based on models derived from range equations on flapping flight (Alerstam and Hedenström 1998) and several detours were evaluated, including a detour to north‐eastern Sahel via Greece for red‐backed shrikes breeding in southern France. Assuming that the red‐backed shrikes were migrating in short hops before fattening up for the barrier crossing over the Mediterranean and Sahara, the extra detour was found to be advantageous over a direct crossing regarding travel costs (Alerstam 2001). Nevertheless, a proper evaluation of this detour requires detailed knowledge of the individual spatiotemporal schedules followed when detouring and crossing. Erni et al. (2005) showed winds to be equally important for optimal detours.

Continuous advances in tracking technology now enable us to track individuals, allowing us to test theoretical predictions using the framework developed by Alerstam (2001) on observed tracks. Here, we use empirical data collected by direct tracking of red‐backed shrikes using archival light‐level loggers (geolocators) from a breeding site in north‐western Spain to evaluate whether the original predictions hold when considering individual routes compared to a direct route toward staging sites in the Sahel and a southern detour with direct crossing of the Sahara from breeding grounds. In addition, we evaluate the potential effects of wind assistance along these routes taking individual timing of migration into account.