The relative fitness of individuals across a population can shape distributions and drive population growth rates. Migratory species often winter over large geographic ranges, and individuals in different locations experience very different environmental conditions, including different migration costs, which can potentially create fitness inequalities. Here we used energetics models to quantify the trade-offs experienced by a migratory shorebird species at locations throughout the nonbreeding range, and the associated consequences for migratory performance, survival, and breeding habitat quality. Individuals experiencing more favorable winter conditions had higher survival rates, arrived on the breeding grounds earlier, and occupied better quality breeding areas, even when migration costs are substantially higher, than individuals from locations where the energy balance on the wintering grounds was less favorable. The energy costs and benefits of occupying different winter locations can therefore create fitness inequalities which can shape the distribution and population-wide demography of migratory species.
Species distribution ranges and demography are constantly shaped by the processes influencing the fitness of individuals occupying different locations. For many migratory species, distribution ranges span large geographic areas, encompassing very different environmental conditions (Newton 2008). Large-scale variation in local weather conditions (Castro et al. 1992, Hötker 2002) and in the quality and quantity of food resources (van Gils et al. 2005, Mathot et al. 2007) can result in differing energy costs and benefits across the range (Quaintenne et al. 2011). The conditions experienced during the winter season can also carry over into subsequent stages of the annual cycle (Gill et al. 2001a, Norris et al. 2004, Gunnarsson et al. 2005a). Carryover effects can operate through individual body condition, with individuals that experience favorable conditions during winter attaining higher body condition for subsequent migration and breeding (Marra et al. 1998, Smith and Moore 2003, Bearhop et al. 2004). Variation in environmental conditions across the winter range can therefore drive variation in individual survival, subsequent breeding success, or both. However, the costs and benefits of different winter locations will also be influenced by distances from the breeding grounds and consequent costs of migration, as higher living costs at sites closer to the breeding areas may be offset by lower energy and mortality costs resulting from a shorter migration.
Migratory shorebirds typically winter along the coastal edges of continents over large latitudinal scales, across which individuals encounter very different environmental conditions (Castro et al. 1992, Piersma et al. 1993). Living costs are often lower for individuals wintering at lower latitudes, mostly as a consequence of local climatic conditions influencing thermoregulatory expenditure (Drent and Piersma 1990, Castro et al. 1992). However, harsher weather conditions in northern wintering areas can be coupled with higher prey availability (van de Kam et al. 2004, Mathot et al. 2007), which can result in similar costs and benefits of occupying winter sites across a broad latitudinal range (e.g., Piersma et al. 1991, Castro et al. 1992). The population size, distribution, and range of migratory species can therefore be determined by the magnitude and direction of the trade-offs associated with each winter location, and their consequences for individual fitness (Newton 2004, Harrison et al. 2011).
Icelandic Black-tailed Godwits Limosa limosa islandica breed almost exclusively in Iceland and winter in coastal areas of Western Europe, from Britain and Ireland (∼56°N) in the north to the Iberian Peninsula (∼36°N) in the south (Gill et al. 2002). As with most shorebird species, adult Icelandic godwits are highly site faithful in summer and winter (Gill et al. 2002). The winter location of ∼860 individuals is currently known, as a result of a long-term marking and tracking program in which ∼1–2% of the population is individually color-ringed and tracked by a network of >2000 volunteers across Europe (Gunnarsson et al. 2006). Previous studies of Icelandic godwits have shown that early arrival is positively related to enhanced breeding success (Gunnarsson et al. 2005a, 2006), and that individuals wintering in good quality habitats also tend to occupy good quality breeding habitats, while individuals from poorer quality winter habitats tend to occupy poor-quality breeding habitats (Gill et al. 2001a, Gunnarsson et al. 2005a). In addition, over the last century this population has expanded into areas of poorer quality habitat in both breeding and nonbreeding ranges (Gunnarsson et al. 2005b), and males in poorer quality breeding sites are more likely to be unpaired (Gunnarsson et al. 2012). Individuals in traditionally occupied, high-quality areas are therefore likely to experience higher breeding success than those in more recently occupied, poorer quality areas. This seasonal matching of habitat quality and fitness may be a major driver of population dynamics in migratory systems, and is likely to have shaped the rate and pattern of population expansion throughout this migratory range. As individuals from the furthest winter locations can arrive in Iceland earlier than those from locations closer to the breeding grounds (Alves et al. 2012a), something other than migration distance must be influencing the timing of arrival in this system. However, the mechanisms driving seasonal matching and, in particular, the way that variation in migration costs and environmental conditions across the winter range impact upon fitness remain unknown.
In order to assess the role of winter environmental conditions in shaping the variation in individual fitness throughout a migratory population, we first quantified the net energy input for Icelandic godwits wintering in three different locations that encompass the entire range and, between them, support ∼40% of the godwit population in winter (Gunnarsson et al. 2005c). We then estimated the energy thermoregulatory costs (maintenance metabolism) of individuals wintering in these locations by adapting a shorebird model that incorporates local weather variables. The consequences of wintering in different locations were then assessed by relating these energy trade-offs to individuals' migratory performance, annual survival rates at these locations, and to their probability of breeding in areas colonized at different periods during the population expansion.
Foraging behavior, time budgets, and energy intake rates
Three locations that span the entire latitudinal range of Icelandic godwits in winter (Gill et al. 2002, Gunnarsson et al. 2005c) are compared: east England, south Ireland, and west Portugal (Fig. 1). Detailed measurements of godwit intake rates and time budgets (see Appendix A) were used to calculate monthly (October to March) net energy intake rate (NEIR) at each winter location using the following formula:
where IR is the average instantaneous intake rate (ash-free dry mass [AFDM] in g/h) and FP is the average daily foraging period (h) for each month. EP is the energy content of prey (kJ/g) and AE is the assimilation efficiency (%), both of which are habitat specific (hab) and derived from published literature (Appendix: Table A2). For locations where more than one foraging habitat is used, NEIR was calculated separately for each habitat and summed for all habitats (summed FP from all habitats never exceeded 24 h). Foraging during nocturnal low-tide periods was both excluded and included from the estimations, providing a range of NEIR for each location.
Energy demands at locations throughout the winter range
To estimate energy demands at each location, we parameterized the model developed by Wiersma and Piersma (1994) for Icelandic godwits. This model estimates maintenance metabolism (Mmaint) in accordance with the basal metabolic rate (BMR), body height, and thermal conductance (TC) of a species, and the thermoregulatory costs associated with different habitats and local weather conditions (see Appendix).
Linking winter location with survival and migratory and breeding performance
Long-term marking and tracking of individual Icelandic godwits has been carried out throughout the entire migratory range since 1995 (Gill et al. 2001a, Gunnarsson et al. 2004, Alves et al. 2012a). Estimates of annual adult survival rates for godwits marked in east England between 1995 and 2000, the period encompassing energetics measurements in this location, have previously been published (Gill et al. 2001a). Annual adult survival rates for Icelandic godwits wintering in west Portugal and south Ireland in 2006–2008 were calculated using Cormack-Jolly-Seber models (see Appendix A) in program MARK (White and Burnham 2000), and differences in survival rates between locations were assessed by 95% confidence interval overlap. The energy costs of migratory flights from each winter location to Iceland were estimated using Flight model (Version 1.18; Pennycuick 2008), parameterized for Icelandic godwits (details in Alves et al. 2012a).
Godwits arriving in Iceland during late April and early May congregate on a few sites that have been regularly monitored every spring since 1999 in order to determine arrival dates of marked individuals (details in Gunnarsson et al. 2006). Individual arrival dates are highly repeatable between years (Gunnarsson et al. 2006), and thus, average arrival dates for individuals from each wintering location were calculated. In addition, marked individuals have been recorded breeding throughout Iceland, and each of these individuals was assigned the year of colonization of the region within which its breeding territory was located (details in Gunnarsson et al. 2005b). Colonization years of breeding regions occupied by birds from each winter location were then compared, as traditionally occupied regions are, on average, higher quality breeding habitats (Gunnarsson et al. 2005b).
Energy balance and spring migration costs from distinct winter locations
Icelandic godwits wintering in west Portugal throughout 2006–2008 never experienced daily maintenance metabolism requirements in excess of BMR (BMR = 173.7 kJ/day) and, for godwits wintering in south Ireland, daily Mmaint only exceeded BMR on 10% of days (Mmaint = 173.9 ± 0.1 kJ/d [all values shown are mean ± SE]). However, in east England, maintenance metabolism requirements exceeded BMR on 92% of days (Mmaint = 212.1 ± 7.0 kJ/d). Godwits wintering in west Portugal do not therefore need to spend any energy in thermoregulation and, for godwits wintering in south Ireland, the amount of energy required for thermoregulation is negligible, as the maximum daily Mmaint ever recorded during these winters (183.2 kJ/d) is still very close to BMR. In east England, however, energetic expenditure for thermoregulation would have been necessary in all winter months (Mmaint > BMR; Fig. 2a) and, in January and March, these demands exceeded the recorded energy input.
In west Portugal, Icelandic godwits can achieve a monthly net energy intake rate (NEIR) of, on average, 1 to 1.5 times as great as the NEIR recorded in south Ireland and 1.3 to 2 times as great as that recorded in east England, depending on whether they forage during one or both tides per day (Fig. 2b). As costs exceed minimum NEIR in east England in all months, godwits wintering in east England must feed during both low tides, whereas in west Portugal or south Ireland godwits can meet their thermoregulatory costs by feeding only during one low tide in a 24-h period (Fig. 2b).
The high NEIR in west Portugal, coupled with lower thermoregulatory costs, allows these Icelandic godwits to potentially experience a positive mean energy balance of 244.3 ± 17.7 kJ/d across the winter months, if they foraged on both tides (Fig. 3a). This average energy surplus (the difference between energetic input [NEIR] and energy costs [Mmaint]) is significantly lower in south Ireland (106.1 ± 12.2 kJ/d), and in east England (−0.6 ±13.9 kJ/d; Kruskal-Wallis test, H2 = 15.16, P = 0.001; Fig. 3a). During spring migration, the majority of Icelandic godwits wintering in Portugal undertake two flights in order to reach Iceland (Alves et al. 2012a), with the majority of them stopping over in the Netherlands (covering a total migratory distance of ∼3800 km), whereas most Icelandic godwits wintering in south Ireland and east England are able to reach Iceland in one flight (covering ∼1425 km and ∼1560 km in total, respectively). The migratory flight to Iceland for a male godwit of average mass departing from south Ireland or east England requires 1290 kJ and 1400 kJ, respectively, while departure from west Portugal requires a total of 2790 kJ, of which 1540 kJ are required for the first migration to the stopover sites (Fig. 3b). Assuming that the net energy intake above Mmaint during March (when migration is initiated) could be converted into available flight energy, godwits wintering in south Ireland require ∼10.5 days of foraging to meet their flight demands, while godwits in west Portugal require ∼8 days to attain the energy needed to reach the stopover locations and ∼14.4 days to cover the entire distance to Iceland in a single flight. For godwits in east England, the net energy intake was, on average, below Mmaint during March of these years; hence, these godwits are likely to have required supplementary foraging locations before departure to be able to meet the energy needs of completing the journey to Iceland.
Arrival dates and colonization year in Iceland and annual survival across the wintering range
Arrival dates in Iceland of marked individuals indicate that godwits wintering in east England arrive significantly later than godwits wintering in south Ireland or west Portugal (F2, 120 = 14.8, P < 0.001; Fig. 3d). In addition, godwits have expanded throughout lowland Iceland over the last century, and more recently occupied areas have a higher proportion of poorer quality breeding habitat (Gunnarsson et al. 2005b). The mean date of colonization of areas occupied by individuals wintering in west Portugal and south Ireland is significantly earlier than for individuals wintering in east England (F2,95 = 3.12, P < 0.05; Fig. 3e), indicating that birds from east England are more likely to breed in poorer quality locations. Finally, average annual survival rates of adult Icelandic godwits marked in east England were lower than for those wintering in west Portugal and south Ireland (Fig. 3c). The confidence intervals of these survival rates overlap, but they translate into estimated average lifespans of ∼7.2 years (east England), and ∼10.6 years (west Portugal and south Ireland).
The distribution and demography of migratory species are constantly shaped by the relative fitness experienced by individuals throughout the range. In Icelandic godwits, high energy costs experienced in winter are associated with lower survival, delayed arrival in Iceland (despite a less costly migration), and a lower probability of occupying good quality breeding habitat. As energy costs typically increase with latitude in winter, occupying costly winter locations could be traded-off with subsequent benefits of a shorter migration. However, in this system, the energy benefits of the more distant winter location are sufficiently positive to outweigh substantially greater migration costs, and to facilitate early arrival and use of better quality breeding habitat. Thus, the environmental conditions and overall energetic balance experienced during winter can produce population-wide fitness inequalities in both the breeding and nonbreeding seasons.
Trade-offs associated with different winter locations
Icelandic godwits wintering in west Portugal not only experience higher energy intake rates, but also have considerably lower thermoregulatory costs than conspecifics wintering in south Ireland and east England. The higher energy intake rates attained by godwits in west Portugal are likely to be primarily a consequence of higher prey biomass. Whereas bivalve densities in east England seldom exceed 1000 individuals/m2 (Gill et al. 2001b), and high densities are only recorded during the first weeks of winter, densities above this threshold are regularly found in west Portugal throughout the winter (Lourenço et al. 2005, Alves et al. 2012b). Moreover, while godwits in west Portugal forage mostly on bivalves above 10 mm in length and on polychaetes above 20 mm (Moreira 1994), prey of these large sizes and biomass are rarely captured in east England or south Ireland (Gill et al. 2001b, Hayhow 2009).
The high NEIR achieved by godwits in west Portugal allows the thermoregulatory energy demands to be met by foraging for only ∼5.1 h/d on the mudflats. Previous studies of godwits in west Portugal also concluded that godwits need only forage on the diurnal low-tide period (Moreira 1994, Lourenço et al. 2008). By contrast, in south Ireland, foraging for 9 h/d (the maximum available time on the tidal mudflats) would be insufficient to cover thermoregulation, and additional foraging on grasslands is required (and indeed accounts for ∼67% of net energy intake rate). In east England, despite godwits gradually increasing their foraging time over the winter, Mmaint was predicted not to be met in both January and March (Fig. 2b). Rapid overwinter depletion of prey resources on mudflats in east England (Gill et al. 2001b) explains the lower net energy intake during this period. The average monthly instantaneous intake rates on mudflats in south Ireland and east England are similar (in south Ireland, 0.27 ± 0.01 mg AFDM/s; in east England, 0.30 ± 0.03 mg AFDM/s, U12 = 8.0, Z = −1.61, P = 0.13), thus, the difference in energy input between these two sites is primarily due to the ability of godwits in south Ireland to use coastal grasslands as foraging locations throughout the winter (average instantaneous intake rates on grasslands, 0.40 mg AFDM/s ± 0.05; Hayhow 2009).
Warmer air temperatures and considerably lower wind speeds also mean that thermoregulatory costs are lower for godwits wintering in west Portugal than in south Ireland or east England. In addition to regional differences in weather conditions, several of the estuaries in east England are large and open in structure, whereas the estuaries in south Ireland tend to be more enclosed and sheltered, which is likely to exacerbate differences in wind speed. Lower energy costs of living at more southerly latitudes has been described for other migratory species (Drent and Piersma 1990, Castro et al. 1992, Wiersma and Piersma 1994) and, in some cases, this accords with individuals at southerly latitudes having lower basal metabolic rates (Kersten et al. 1998). However, BMR is unlikely to vary considerably among temperate locations, and internal regulation of BMR would be likely to exacerbate the differences in energy balance between northern and southern locations, as a lower BMR would be expected for the warmer locations (Castro et al. 1992, Kersten et al. 1998). Thus, the variation in costs of living for Icelandic godwits in different winter locations are likely to result primarily from the influence of local weather conditions on the energy costs of thermoregulation.
Carryover effects and population-wide implications of winter site choice
Icelandic godwits wintering in west Portugal enjoy a more positive energetic balance and arrive in Iceland significantly earlier than godwits wintering in east England, despite undertaking a spring migration that demands double the energy. The very positive energetic balance available at this location is therefore sufficient to both fuel the longer migration and facilitate early arrival (Alves et al. 2012a). Wind assistance can reduce the costs of migration (Shamoun-Baranes et al. 2010) and, in other migratory systems, the higher costs of migration for individuals wintering further from the breeding grounds can be offset by wind subsidies encountered en route (Piersma et al. 1991). Although such wind subsidies might influence very long-distance flights over different routes, Icelandic godwits undertake the final part of their journey along the same route independently of winter location (Alves et al. 2012a), and wind subsidies are thus likely to affect all individuals during the sea crossing at the last leg of the journey.
The variation in energy surpluses across the winter range is likely to also influence survival. In fact, NEIR in east England is predicted to have fallen below Mmaint in some months during the winters of 1995–1997 (Fig. 2b), and lower survival rates during these winters (Fig. 3a) are likely to be related to such energy deficits. For Icelandic godwits wintering in west Portugal and south Ireland, survival rates are estimated to be higher and similar, and indeed at both sites NEIR provides a favorable energy balance.
In avian systems, breeding success has been positively correlated with arrival dates, with birds occupying the best quality breeding sites often arriving earlier than those in the poor-quality breeding sites (Marra et al. 1998, Smith and Moore 2003, Gunnarsson et al. 2005a). Thus, the energy advantages gained by wintering in west Portugal and south Ireland are likely to be translated into higher breeding success, through early arrival in Iceland and possibly also through individuals being in better condition for breeding. Godwits from these winter locations are also likely to breed in traditionally occupied areas with a greater relative abundance of good quality breeding habitat, thus wintering in energetically favorable locations is associated with breeding early and in better quality habitats. Differences in the energy trade-offs that occur at different winter locations may therefore be a key component of the strong seasonal matching of habitat quality and fitness in this migratory system (Gunnarsson et al. 2005a), and of the role this can play in population regulation (Gill et al. 2001a).
Many migratory species show strong seasonal connectivity, and greater survival and breeding success of individuals wintering in more favorable locations means that these individuals will contribute disproportionately to the overall population demography (Gunnarsson et al. 2005a). This is particularly relevant for long-lived and highly philopatric species as, once settlement occurs, individuals are likely to experience similar breeding and winter conditions throughout their lives (e.g., Lok et al. 2011). If some winter locations are consistently more advantageous than others, then selection of winter location by new recruits is likely to be of paramount importance for individual fitness and population growth rates. Individuals wintering in less energetically favorable locations may also be more constrained in their capacity to explore alternative sites, hence reinforcing winter site fidelity. The range size, demography, and population dynamics of migratory species can therefore be strongly influenced by winter site selection and the subsequent environmental conditions and trade-offs experienced by individuals across the entire range.
We are very thankful to the many hundreds of volunteers involved with catching, ringing, and reporting of color-ringed godwits, especially the Farlington and Wash Wader Ringing Groups. We thank Catriona Morrison and Phil Atkinson for help with data analysis and three anonymous reviewers for their comments on an earlier version of the manuscript. This work was funded by the Calouste Gulbenkian Foundation (J. A. Alves), the Arcadia Fund (W. J. Sutherland), and NERC (D. B. Hayhow, J. A. Alves, and J. A. Gill).