So far, so good… Similar fitness consequences and overall energetic costs for short and long-distance migrants in a seabird

Abstract

Although there is a consensus about the evolutionary drivers of animal migration, considerable work is necessary to identify the mechanisms that underlie the great variety of strategies observed in nature. The study of differential migration offers unique opportunities to identify such mechanisms and allows comparisons of the costs and benefits of migration. The purpose of this study was to compare the characteristics of short and long-distance migrations, and fitness consequences, in a long-lived seabird species. We combined demographic monitoring (survival, phenology, hatching success) of 58 Northern Gannets (Morus bassanus) breeding on Bonaventure Island (Canada) and biologging technology (Global Location Sensor or GLS loggers) to estimate activity and energy budgets during the non-breeding period for three different migration strategies: to the Gulf of Mexico (GM), southeast (SE) or northeast (NE) Atlantic coast of the U.S. Survival, timing of arrival at the colony and hatching success are similar for short (NE, SE) and long-distance (GM) migrants. Despite similar fitness consequences, we found, as expected, that the overall energetic cost of migration is higher for long-distance migrants, although the daily cost during migration was similar between strategies. In contrast, daily maintenance and thermoregulation costs were lower for GM migrants in winter, where sea-surface temperature of the GM is 4-7o C warmer than SE and NE. In addition, GM migrants tend to fly 30 min less per day in their wintering area than other migrants. Considering lower foraging effort and lower thermoregulation costs during winter for long-distance migrants, this suggests that the energetic benefits during the winter of foraging in the GM outweigh any negative consequences of the longer-distance migration. These results support the notion that the costs and benefits of short and long-distance migration is broadly equal on an annual basis, i.e. there are no apparent carry-over effects in this long-lived bird species, probably because of the favourable conditions in the furthest wintering area.

Fig 1. Example of long-term monitoring of daily activity patterns and temperature recorded using a combined Global Location Sensor (GLS), immersion and temperature logger attached for 260 days on a Northern Gannet nesting on Bonaventure Island and wintering off the northeast coast of the U.S. Diamonds (◆) depict the wet time (h), when the gannet is diving or sitting on the water; squares (□) depict the dry time (h), when the gannet is out of the water, flying or on land at the colony; and, triangles (Δ) are the mean sea surface temperature per day (± SD,°C).

Breeding periods are bounded by dry time per day of more than 12 h, and the date of arrival at the wintering area is based on geolocation using the light data (see Methods).

Fig 2. Centroid of wintering location (Dec. 15^th–Feb. 15^th, mean ± 95% CI) of 58 Northern Gannets nesting on Bonaventure Island, Canada, equipped with a Global Location Sensor (GLS) logger and tracked in 2010–2011 or in 2011–2012.

Red dots presented individuals wintering in the Gulf of Mexico, yellow dots in southeast coast of the USA and blue dots in northeast coast of the USA. The map was generated in R (R Core Team, 2016) using the ‘ggmap’ [48] and ‘ggplot2’ packages [49] with map tiles by Stamen Design (www.stamen.com) and data by OpenStreetMap, under ODbL, under CC BY 3.0 (creativecommons.org/licenses/by/3.0/).

Fig 3. Annual survival rates of Northern Gannets nesting on Bonaventure Island, Canada, and wintering in the Gulf of Mexico (GM), southeast coast of the U.S. (SE), northeast coast of the U.S. (NE) and all individuals together from 2009 to 2017.

Survival rates for GM, SE and NE were calculated from the model ɸ_g*t p_t. and recapture probability (p) were calculated from the best fit model: ɸ_t p_t (see Table 2). Lines for sites are presented separately just to display the absence of difference, but they are faded because wintering sites are not in the best model (see Table 2). Bars are showing the 95% CI (estimated from profile likelihood method) around the mean annual survival.

Fig 4. Multi-year hatching success (%; number of chicks hatched/eggs laid) of three migrant groups of Northern Gannets nesting on Bonaventure Island, Canada, and wintering in the Gulf of Mexico (GM), southeast coast of the U.S. (SE) and northeast coast of the U.S. (NE) from 2009 to 2017.

There is no interaction between migrant groups and hatching success (P > 0.05) but an interaction between year and hatching success (P < 0.0001). Bottom line n represent the total of nests for all categories. We used only the data for instrumented birds in 2010–2011 and 2011–2012, and for which we monitored hatching success between 2009 and 2017. Given the very high fidelity to wintering sites in gannets [25], we assume that individuals keep the same area year after year.

Fig 5

a) Duration of the non-breeding period (migrations and winter seasons) and b) time spent flying per day of three migrant groups of Northern Gannets nesting on Bonaventure Island, Canada. Wintering areas: Gulf of Mexico (GM, n = 12), southeast coast of the U.S. (SE, n = 18) and northeast coast of the U.S. (NE, n = 28) in 2010–2011 and 2011–2012. (*: P < 0.05; **: P ≤ 0.01; ***: P ≤ 0.001).

Fig 6. Mean sea surface temperature (SST) recorded by loggers on Northern Gannets nesting on Bonaventure Island, Canada, throughout the annual cycle.

Wintering areas: Gulf of Mexico (GM, n = 12), southeast coast of the U.S. (SE, n = 18) and northeast coast of the U.S. (NE, n = 28) in 2010–2011 and 2011–2012.

Fig 7. Comparison of energy expended during the non-breeding period (fall migration, winter and spring migration) between three migrant groups of Northern Gannets nesting on Bonaventure Island.

Wintering areas: Gulf of Mexico (GM, n = 12), southeast coast of the U.S. (SE, n = 18) and northeast coast of the U.S. (NE, n = 28) in 2010–2011 and 2011–2012. a) Energy spent flying per day, b) Energy spent resting on water per day, c) Energy spent flying and resting on water per day (*: P < 0.05; **: P ≤ 0.01; ***: P ≤ 0.001).

Fig 8. Comparison of total energy spent during the non-breeding period (fall migration, winter and spring migration) between three migrant groups of Northern Gannets nesting on Bonaventure Island, Canada.

Wintering areas: Gulf of Mexico (GM, n = 12), southeast coast of the U.S. (SE, n = 18) and northeast coast of the U.S. (NE, n = 28) in 2010–2011 and 2011–2012. Cumulative energy spent for resting metabolic were calculated by multiplying energy spent resting on water per day and duration of the non-breeding period. Cumulative energy spent flying during winter were calculated by multiplying energy spent flying per day during winter and duration of the winter season. Cumulative energy spent flying during migration were calculated by multiplying energy spent flying per day during migration and duration of both migrations (spring and fall).