An annual cycle perspective on energetics and locomotion of migratory animals

Whether you have observed thousands of birds flying overhead, seemingly endless trails of grazers reshaping the Serengeti, rivers swelling with fish moving to spawning grounds or the return of a species heralding the onset of spring, migration is an awe inspiring phenomena in nature. Migratory behaviour is a life history strategy observed across the animal kingdom, anywhere on earth. Migration is considered a specialized behaviour, characterized by persistent and straightened out movement in which station-keeping behaviour (movements within a home range) is temporarily inhibited and specific organization of energy allocation may be needed to support these movements (Dingle, 1996). That migration is considered energetically costly is due to the investment in persistent locomotion required to travel long distances, but whether indeed this is one of the most energetically costly activities within the life cycle of an animal remains to be proved (Carey, 1996).

Animals have evolved diverse behavioural and physiological strategies to meet the energetic demands of migration (Dingle, 1996; Sapir et al., 2011). Movements can be powered entirely by an animal's investment in locomotion or assisted by the environment. For example, birds and insects can use atmospheric flows (Chapman et al., 2011; Gatehouse, 1994; Shamoun-Baranes et al., 2017b) and aquatic species can use currents (Gaspar et al., 2006; Putman, 2018) to save time or energy during long-distance movements. Migrants may also differ in how they organize the storage and allocation of energy to fuel migration bouts. An energetics framework has been proposed, analogous to one used for avian reproduction, to describe and compare different migration strategies along a continuum (Evans and Bearhop, 2022). In this framework, capital migrants may spend days to weeks increasing their energy reserves to fuel extreme endurance migrations moving for days without stopping to rest (Battley et al., 2012; Carneiro et al., 2024; Hedlund et al., 2021). At the other end of the spectrum, income migrants may move for short durations interspersed with (short) pauses for refuelling, or even refuel while travelling (Amélineau et al., 2021; Imlay et al., 2020; Sawyer and Kauffman, 2011). Many species, however, may exhibit intermediate strategies to organize their energy expenditure during migration, and be flexible in how they manage energy expenditure (Halsey et al., 2019).

While persistent movements during migration are potentially energetically costly, migration is only part of the annual life cycle of an animal. For animals living in seasonal environments, life history stages such as migration, breeding and moult may be organized in an annual routine, with each stage occurring in different environments and incurring different energetic costs (McNamara and Houston, 2008b; Wingfield, 2008). Annual cycles will pose temporal and spatial constraints on animals through endogenous and external factors. The duration and timing of reproduction may be highly conserved and under strong endogenous control (Sharp, 2005; Stearns, 1992). Animals will also be constrained in space because of their movement capacity, central place foraging strategies or specific habitat requirements during different stages in the annual cycle. Thus, being in the right place at the right time may be important for survival or reproduction (Condit et al., 2022). Furthermore, migratory animals may exhibit site fidelity to breeding areas as well as non-breeding areas (Brown et al., 2021; Picardi et al., 2023), further constraining their movements. An annual cycle perspective can thus improve our understanding of different life history trade-offs, seasonal interactions and activity-related fluxes in energy expenditure and provide a more integrated approach to animal ecology (Marra et al., 2015; McNamara and Houston, 2008a; see also papers in this Special Issue). An annual cycle approach is important, especially in migratory animals, to identify where and when the main energetic bottlenecks may arise for a species (Marra et al., 2015). In this Commentary, we reiterate earlier calls for an annual cycle perspective. In light of the theme of this Special Issue ‘Integrating biomechanics, energetics and ecology in locomotion’ (Biewener and Wilson, 2025), we focus on daily energy expenditure in migratory animals in the wild, moving in and experiencing variable and heterogenous environments during their annual cycle. Using an example from our research, we demonstrate why studying energy expenditure in the context of an animal's annual cycle is valuable and feasible, in migratory as well as in non-migratory animals.

By incorporating an annual cycle perspective on energetics and locomotion, we can address several questions. (1) How does energy expenditure vary during the annual cycle, and why (Fig. 1)? (2) How does the relative contribution of locomotion to daily energy expenditure vary during the annual cycle (Fig. 2)? (3) At which stage in the annual cycle and where do energetic bottlenecks arise? (4) What behavioural adaptations can animals use to reduce the energetic costs of different (movement) strategies? (5) When and where are animals most susceptible to unpredictable and abrupt environmental change?

Information on the time animals spend on various activities can be used to estimate daily energy expenditure in wild animals (Goldstein, 1988). In certain study systems, or stages of the annual cycle, time–activity budgets can be calculated based on visual observations in the field. Time–activity budgets can then be integrated with field, lab-based or theoretical estimates of the energetic cost of different activities, including variable resting metabolic rates, to convert activity budgets into estimated energy budgets (Dasilva, 1992; Goldstein, 1988; Karasov, 1992). Such methods have been used to estimate energy expenditure throughout the annual cycle in mammals, for example, showing seasonal shifts in time and energy invested in foraging, reproduction and resting (Dasilva, 1992; Kenagy et al., 1989). However, visual observations to measure time–activity budgets are hardly feasible for migratory animals. Thus, without remotely monitoring animal behaviour, data availability is a major limiting factor.

Bio-logging has opened new opportunities for quantifying behaviour and assessing related energy expenditure (Hussey et al., 2015; Kays et al., 2015; Ropert-Coudert and Wilson, 2005; Watanabe and Papastamatiou, 2023). Positional data obtained by bio-logging can be annotated with a behaviour (e.g. using accelerometry) and relevant environmental conditions (e.g. using remote sensing data), and bio-energetic models can then be used to estimate energy expenditure at the temporal resolution of interest (Fig. 1). If information is only available on the position of an animal, its speed can be derived and used to distinguish between stationary periods and locomotion. An average cost of movement based on scaling laws could then be used to estimate the resulting energy expenditure attributed to locomotion (Alexander, 2005) and thermoregulatory costs can be estimated by integrating environmental information. Energetic costs of locomotion could be further refined by using theoretical or empirical relationships between the metabolic cost of locomotion and speed for a given species and mode of locomotion (Alexander, 2003; Taylor et al., 1982; Tobalske et al., 2003). However, annotating positional data with a behaviour based on speed alone is not always sufficient to distinguish between different activities, such as flapping and soaring flight, or resting from walking, which may have contrasting energetic costs (Shamoun-Baranes et al., 2012, 2016). Alternatively, a range of modelling approaches exist in which step lengths and turning angles can be used to estimate behavioural states (Williams et al., 2020). Once behavioural states are classified, the time in each state could be calculated and an average energy expenditure per behavioural state could then be applied.

Behaviours can also be classified, or energy expenditure directly estimated, using additional sensor data. For example, data collected from heart rate loggers (Bishop et al., 2015; Campbell et al., 2008; Grémillet et al., 2005; Seebacher et al., 2005), accelerometers (Collins et al., 2015; Resheff et al., 2014), time-depth recorders (Grémillet et al., 2005; Seebacher et al., 2005) or temperature sensors (Garthe et al., 2012) have been used to estimate energy expenditure. Heart rate loggers can account for various somatic costs such as moult and reproductive development which accelerometers cannot directly account for, whereas accelerometers can be used to distinguish between being stationary and different types of body movement. By converting sensor data into behaviour and incorporating this with the ecology of the study species, we can better understand how energy expenditure and behavioural adaptations are organized within the annual cycle. See Table 1 for several examples of studies in which locomotion and energetics or proxies thereof were studied throughout the annual cycle.

Below we provide some examples from our earlier work (Brown et al., 2023), showing how an integrative approach can be used to study energetics and locomotion it the context of the annual cycle of the lesser black backed gull (Larus fuscus) and briefly address some of the questions posed in the Introduction. Long-term research on the breeding biology, demography and diet of gulls at a breeding colony on Texel, The Netherlands, was complemented with individual tracking data, providing insight into the annual movements (Camphuysen et al., 2024), migratory behaviour (Brown et al., 2021; Shamoun-Baranes et al., 2017a), foraging ecology (Camphuysen et al., 2015; Tyson et al., 2015) and fine-scale flight behaviour in the breeding stage (Sage et al., 2019, 2022; Shamoun-Baranes et al., 2016). In order to study energetics throughout the annual cycle and consequences of different migration strategies, it took additional lab and fieldwork work to calibrate energetic estimates of different activities (Brown et al., 2022a), and several years of tracking data collection from multiple colonies before enough knowledge and suitable data was available to study energy expenditure throughout the annual cycle (Brown et al., 2022b, 2023) (Fig. 1).

Despite breeding in close proximity and having a relatively synchronized breeding stage in the annual cycle (Camphuysen et al., 2024), birds from the same colony had very different migration and overwintering strategies, with destinations ranging from a few hundred (e.g UK) to a few thousand kilometres (e.g. West Africa), either using predominantly urban, agricultural, coastal or marine habitats, and experiencing very different climates, from temperate to tropical (Shamoun-Baranes et al., 2017a). While the average estimated daily energy expenditure (777 kJ day⁻¹) did not differ among these diverse movement strategies, large variations within and among individuals were observed (Fig. 1). For example, during spring migration, a short period of highly elevated daily energy expenditure (1.7× average daily energy expenditure) was observed among the longest distance migrants, and for them it was the stage in the annual cycle in which energy expenditure was highest (Fig. 1), reaching >2000 kJ day⁻¹ in some cases (Fig. 2A). By integrating energetics and locomotion, it is possible to understand the source of these peaks as gulls transition from spending approximately 15% of time in flight per day to spending up to 70% of their time in flight (Fig. 2B). While such a strategy results in elevated energy expenditure, the extended temporal investment in locomotion resulted in a travel distance of more than 1000 km in 24 h, speeding up its return to the breeding colony.

Exploring data from the same individuals over different years can show how flexible they could be in how they organize their time and energy expenditure, for example, in response to weather conditions. Gulls can adopt different flight modes such as energetically costly flapping flight or soaring flight with much lower costs per unit time (Brown et al., 2022a). However, soaring flight requires suitable atmospheric conditions (Sage et al., 2019, 2022; Shepard et al., 2016; van Erp et al., 2023). An example of inter-annual and intra-individual variation in partitioning of energy expenditure during migration from Africa to The Netherlands is provided in Fig. 2B,C. In two consecutive years, the same individual spent 70–75% of time in flight on the same day of the year while migrating back to the breeding ground. Yet in 2018, it spent a larger proportion of time soaring (39% compared with 2%), greatly reducing daily energy expenditure (Fig. 2C). While crossing the Bay of Biscay, the gull spent 29% of its time soaring, with an estimated daily energy expenditure that was similar to the mean estimated daily energy expenditure for gulls throughout the entire annual cycle. These examples show how fine scale flexibility in flight behaviour and opportunistic response to environmental conditions can have important implications for daily energy expenditure.

In gulls, the breeding stage can be an extended period of elevated energy expenditure (Fig. 1) because of a relatively high investment in locomotion. During this stage, both males and females incubate eggs and care for their young, regardless of migration strategy. Time–activity budgets also change during the breeding season, for example, travel distances per foraging trip are higher during the chick-rearing stage compared with during incubation (Camphuysen et al., 2015), and adults spend less time at the nest and more time in flight (Spelt et al., 2019) as they need to acquire food for themselves and their young. While daily energy expenditure was generally above the annual mean for all individuals during the breeding season, for birds with relatively short migration distances, the breeding season tended to be the stage in the annual cycle with the highest estimated daily energy expenditure (1.4× average daily energy expenditure, Fig. 1). Despite such differences, on average, migration strategies had no impact on reproduction success or on adult survival (Kentie et al., 2023). Regardless of the strategy, gulls maintained levels of activity throughout the annual cycle below maximum theoretical thresholds considered sustainable (Drent and Daan, 1980; Dunn et al., 2020), suggesting that energetic bottlenecks may be quite rare in the study population.

While we advocate an annual cycle perspective where possible, we acknowledge that for some species, studying energetics and locomotion throughout the year is still impossible because of technological limitations. Nevertheless, increasingly small activity loggers are being used to study activity year round, for example, in smaller birds that weigh less than 100 g, which provide promising opportunities to study energetics throughout the annual cycle (Briedis et al., 2020; Wong et al., 2024). Estimates of the cost of locomotion, or other behaviours may still be lacking for quite a few species, but scaling laws (Alexander, 2003) or other theoretical frameworks (e.g. aerodynamic theory; Pennycuick, 1989) may provide reasonable alternative estimates. The computational work needed to annotate movement data with environmental data at the right temporal and spatial scale and estimate daily energy expenditure is also a challenging component of such research. To some extent, artificial intelligence tools and data annotation software to help automate several of these steps will reduce the technical challenges (Chimienti et al., 2022; Dodge et al., 2013; Nathan et al., 2012).

An additional limitation to acquiring data throughout the annual cycle is that several tracking methods still require sensor retrieval or a second surgical intervention to remove implanted tags (Bishop et al., 2015; Macías-Torres et al., 2022). Furthermore, large numbers of individuals may be needed before sufficient data are available to address research questions. As a result, research may also be biased towards, or focus on, successful individuals (e.g. individuals that successfully complete an annual cycle), masking instances when animals die because they have exceeded an energetic ceiling. It would be interesting to compare patterns of daily energy expenditure between complete annual cycles and truncated annual cycles, to better understand potential causes of mortality. Similarly, animals may skip or abandon stages in the annual cycle, for example, birds may forego breeding or animals may enter so called ‘emergency’ life history stages (Spaans et al., 1998; Wingfield, 2003). Comparative analysis among alternative annual cycle structures could improve our understanding of energetic trade-offs among different stages in the annual cycle.

Although different methods exist for estimating daily energy expenditure, energy intake is still very challenging to estimate in free-ranging animals. Thus, studies of energetics and locomotion often focus on costs of resting and locomotion, rather than energy budgets which account for energy intake to better assess energetic benefits of different lifestyles. To some extent, foraging activity and food intake could be estimated from animal-borne sensors such as accelerometers or temperature loggers (Clermont et al., 2021; Grémillet and Plös, 1994; Lok et al., 2023). For parts of an annual cycle, field measurements could provide supplementary information on foraging activity and intake rates (Dasilva, 1992; Ginnett and Demment, 1997; Gloutney et al., 2001). In studies focusing on movement, the energetic costs of stationary activities are likely underestimated, even when accounting for variable thermoregulatory costs due to environmental conditions. For example, moult, lactation and egg production are periods of elevated metabolic costs (Croxall, 1982; Guillemette and Pelletier, 2022; Guillemette et al., 2007; Nilsson and Råberg, 2001; Thometz et al., 2021, 2016). Similarly, somatic growth and immune function also incur energetic costs (Buehler and Piersma, 2008; Piersma et al., 1996) that are not easily accounted for with time–activity budgets. Combining empirical work with theoretical frameworks for modelling dynamic energy budgets in combination with fine scale information on movement (Chimienti et al., 2020) may provide opportunities for better understanding of where and when animals reach energetic limits or surplus.

Annual cycles and thus energy partitioning is likely to differ between sexually immature animals and animals that have reached a reproductive stage. Juveniles may take time before they reach the locomotor (e.g. flight) and foraging efficiency of adults, thus costs of locomotion may be higher during early life stages (Enstipp et al., 2021; Jouma‘a et al., 2024; Rotics et al., 2016; Weston et al., 2018). Their migration routes and timing, habitat requirements and space use may also differ from adults (Sergio et al., 2014; Vansteelant et al., 2017) and individuals will need time to develop effective foraging strategies and dietary specializations (Grecian et al., 2018; Jeglinski et al., 2012). Therefore, tracking studies of animals in their first years of life, will provide great potential for studying how past experience and spatio-temporal constraints in the annual cycle influence energy expenditure.

We hope this Commentary sparks interest in applying an annual cycle perspective when studying energetics and locomotion in the wild. An annual cycle perspective is likely feasible in many species considering the number of species that are tracked for a complete annual cycle (e.g. references in Table 1). The approach we describe will also be highly valuable for non-migratory species. While there are constraints to data collection, we are experiencing a golden age compared with the challenges faced by researchers studying annual cycles of energy expenditure in free-ranging animals just a few decades ago. We now have an opportunity to study relative investments in different activities, compare investments across taxa, or foraging and migration specializations. By combining energy expenditure, behaviour and environmental conditions, we can study the breadth of behavioural responses to variable environments and demands during an annual cycle and their energetic consequences. Hopefully, such an approach can improve our understanding of the capacity of animals to adapt to predictable and unpredictable environmental change.

J.S.-B. thanks participants from the symposium Integrating Biomechanics, Energetics and Ecology in Locomotion for stimulating discussions and The Company of Biologists for their organisation and sponsorship of the symposium. We thank Morgan Brown for earlier discussions and her research on energy expenditure of gulls. Our many gull collaborators have been invaluable in making our long term research possible. UvA-BiTS studies are facilitated by infrastructures for e-Ecology, developed with support of NLeSC and LifeWatch and carried out on the Dutch national e-infrastructure subsidized by the NWO Domain Science (2021.030) with support of the SURF Cooperative. Part of the tracking research presented in this paper were supported by Open Technology Program (17083), which is financed by NWO Domain Applied and Engineering Sciences, in collaboration with the following private and public partners: Rijkswaterstaat and Gemini Windpark. We thank Edwin Baaij and Willem Bouten for their years of support with UvA-BiTS. We thank Stacy Shinneman (IBED, UvA) for preparing the figures.