Balancing Energy Budget in a Central-Place Forager: Which Habitat to Select in a Heterogeneous Environment?

Balancing Energy Budget in a Central-Place Forager: Which Habitat to Select in a Heterogeneous Environment?

Abstract

Foraging animals are influenced by the distribution of food resources and predation risk that both vary in space and time. These constraints likely shape trade-offs involving time, energy, nutrition, and predator avoidance leading to a sequence of locations visited by individuals. According to the marginal-value theorem, a central-place forager must either increase load size or energy content when foraging farther from their central place. Although such a decision rule has the potential to shape movement and habitat selection patterns, few studies have addressed the mechanisms underlying habitat use at the landscape scale. Our objective was therefore to determine how Ring-billed gulls select their foraging habitats while nesting in a colony located in a heterogeneous landscape. Based on locations obtained by fine-scale GPS tracking, we used resource selection functions and residence time analyses to identify habitats selected by gulls for foraging during the incubation and brood rearing periods. We then combined this information to gull survey data, feeding rates, stomach contents, and calorimetric analyses to assess potential trade-offs. Throughout the breeding season, gulls selected landfills and transhipment sites that provided higher mean energy intake than agricultural lands or riparian habitats. They used landfills located farther from the colony where no deterrence program had been implemented but avoided those located closer where deterrence measures took place. On the other hand, gulls selected intensively cultured lands located relatively close to the colony during incubation. The number of gulls was then greater in fields covered by bare soil and peaked during soil preparation and seed sowing, which greatly increase food availability. Breeding Ring-billed gulls thus select habitats according to both their foraging profitability and distance from their nest while accounting for predation risk. This supports the predictions of the marginal-value theorem for central-place foraging over large spatial scales.

Introduction

Animals face time and energy constraints leading to trade-offs in their activity budget, which can also be modulated by factors such as the spatio-temporal distribution of food resources, conspecifics, predation risk, and phenology. How animals respond to these constraints in order to maximize their fitness through foraging behavior has been the main focus of optimal foraging theory. For instance, the marginal-value theorem has been used to predict which resource patch an animal should exploit and how long it should stay before moving to another patch or return to its nest or shelter. Assuming that animals maximize their net energy gain, this model has provided relevant qualitative predictions. However, it has been developed and used for small-scale systems in which animals are assumed to incur few or no travel costs and to be highly informed about their environment.

This model may therefore be difficult to apply at the landscape level because of information uncertainty about the environment, which influences learning ability and because of the limited motion and navigation capacity of animals. For example, classical central-place foraging models based on the marginal-value theorem predict that prey load size should increase with the distance traveled by a forager from its central place. However, a forager moving across the landscape with a large load can incur increased travel costs due to greater energy expenditures or can encounter higher predation risks through increased exposure and reduced maneuverability. Therefore, the impact of carrying a heavy load can influence the time and energy budget of a central-place forager in different ways, sometime far from the conclusions of the classical models.

The marginal-value theorem predicts that a foraging path is the outcome of balancing trade-offs between energy expenditures and gains, especially within landscapes where resources are heterogeneously distributed. Although it is difficult to use the marginal-value theorem to make precise predictions under relaxed assumptions, classical central-place foraging models nevertheless allow to predict that distant patches must provide higher energy "prey" than those found in nearby patches. Hence, the profitability of a given load size may vary for a generalist forager traveling through a heterogeneous landscape. Also, habitats providing low energy food should only be used close to the central place whereas habitats with high-energy food may be exploited near or far from the central place. It remains that travel costs may increase the use of poor quality habitats when individuals must sample and learn the quality of their environment. Moreover, temporal variation in habitat availability and forager condition may alter the pattern of habitat use along a distance gradient.

Although assessing the costs and benefits of large spatio-temporal scale movements is difficult, analytical methods based on accurate location data (e.g., GPS) are now available to study movement behavior. Combining these analytical methods with in situ observations of individual foraging strategies, patch quality, and environmental conditions while considering the individuals' characteristics has the capacity to provide insights into the cost-benefit trade-offs associated with foraging movements underlying habitat selection. For instance, resource selection functions have been widely used to assess habitat selection. They are based on the comparison of relative habitat use (defined by presence-only data) and availability or on the presence/absence of individuals in habitat patches. Resource selection functions are particularly informative if a distinction can be made between actively selected locations, such as foraging patches, and the incidentally selected locations visited during inter-patch movements. Bastille-Rousseau et al. have advocated the use of a combination of resource selection functions, residence time analysis, and ground surveys to study resource selection and foraging strategies at the landscape level. Considering the hierarchical aspect of the selection process, the difficulty of defining available habitats with presence-only data can be avoided by building resource selection functions based on the habitats actually visited for foraging vs. those crossed when moving to a patch. Measuring the time spent by an animal within the surroundings of recorded locations (residence time) should allow discriminating between locations occurring within foraging patches and those found along movement paths.

We used resource selection functions and residence time analyses from GPS-tracking data, as well as survey data, feeding rates, stomach contents, and calorimetric analyses to study the processes that determine habitat use by breeding Ring-billed gulls. This species is a colonial central-place forager that feeds opportunistically upon a wide variety of prey items found in both aquatic and terrestrial habitats. We expected that gulls should be more likely to forage in a patch where the amount of habitats providing high-energy food increases and that such a relationship should be more pronounced far from the colony so that gulls reach a threshold of profitability. We also hypothesized that gulls should select habitats with a temporally variable food availability only when those habitats provide high food returns. For instance, agricultural lands and lawns should be selected on rainy days when annelids (earthworms) are more available to gulls. By testing these predictions, our study sheds light on the process of habitat selection by animals from an energy trade-off perspective.