Determinants of false alarms in staging flocks of semipalmated sandpipers

Guy Beauchamp 0 1 2 0 The Author 2010. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions , please 1 @umontreal.ca. Received 10 September 2009; revised 17 December 2009; accepted 12 February 2010 2 Faculty of Veterinary Medicine, University of Montre al , PO Box 5000, St-Hyacinthe, Que bec , Canada J2S 7C6 False alarms occur when animals flee abruptly upon detection of a threat that subsequently proved harmless. False alarms are common in many species of birds and mammals and account for a surprisingly high proportion of all alarms. False alarms are expected to be more frequent in larger groups, where the odds of misclassifying threats are higher, and under environmental conditions where detection of threats is compromised, such as low light levels. In addition, false alarms should be less frequent when the energetic cost of fleeing increases. I examined these hypotheses in roosting flocks of staging semipalmated sandpipers (Calidris pusilla) over 2 years. False alarms increased with group size but the effect of group size was confounded by the fact that more attacks by falcons (Falco spp.) were directed at larger roosts. False alarms were more frequent at low light levels and later during staging. As individuals double their body mass during staging, the energetic cost of fleeing must greatly increase thus contributing to decreased responsiveness. A simple reduction in responsiveness caused by repeated exposures to harmless signals would also produce a temporal decrease in responsiveness but this hypothesis cannot account for the effect of group size and light level. Study of the determinants of false alarms provides an opportunity to examine adjustments in behavior in relation to changes in perceived predation risk. Key words: false alarm, group size, predation risk, roosting, semipalmated sandpiper, stopover ecology. [Behav Ecol 21:584-587 (2010)] - Aupon detecting potential threats. Such threats, however, nimals often interrupt their activities and flee to cover often prove harmless and much time is thus spent fleeing at the expense of fitness-enhancing activities such as feeding. False alarms are a common feature in many species of birds and mammals and account for a surprisingly large proportion of alarms in some species (Hoogland 1981; Cresswell et al. 2000; Blumstein et al. 2004; Kahlert 2006). It is thought that a high level of responsiveness to potential threats is a low-cost strategy that reduces the likelihood of not responding appropriately when a real attack does occur, the better-safethan-sorry principle. For a single individual, the appropriate level of responsiveness should depend on the costs and benefits of responding or not responding to potential threats whether they prove real or harmless. The costs of responding include the energetic costs imposed by fleeing and the foregone returns from any curtailed activity. Not responding upon detection of a potential threat allows foragers to avoid these costs but at the risk of having to deal with a real predator. Not responding is also more dangerous if threat detection is hampered. This can occur when detection conditions are not ideal, for instance at low light levels (Lima 1988), or in the presence of visual obstruction (Guillemain et al. 2001), or if harmless and real threats share many features, making discrimination more difficult. In a group, individuals can rely on their own detection as described above but can also react to the responses of companions (Pulliam 1973). When relying on detection by companions, it is not always obvious to decide whether departures are caused by a real alarm or are due to nonthreatening factors, such as satiation for instance (Lima 1995). However, given that multiple, simultaneous departures are more likely to represent independent responses to a real threat, individuals in groups that have not detected the threat by themselves could rely on these multiple departures to take appropriate action (Lima 1994; Proctor et al. 2001; Beauchamp and Ruxton 2007b). Although the risk of not detecting a potential threat is certainly lower in a group because many more eyes are scanning the surroundings, false alarms can still occur in a group given that the first few detectors are still relying on their own potentially biased assessment of threats and nondetectors rely on this potentially biased information. In fact, given that there are more potential detectors in a group, one would predict that false alarms should actually increase with group size because the odds of misclassifying a threat, at the level of the group, increase with the number of independent assessments of a potential threat (Treisman 1975). Not responding in a group is also costly because individuals that are left behind following the departure of companions may be more at risk of attack by predators (Quinn and Cresswell 2005; Beauchamp and Ruxton 2007a; Sirot and Touzalin 2009). Surprisingly, few studies have examined the determinants of false alarms in animal groups. In a study of foraging redshanks (Tringa totanus), individuals were more likely to take flight in alarm on cloudy days and later in the overwintering season (Hilton et al. 1999). Clouds decrease light levels thus increasing the odds of misclassifying threats. Proneness to alarm later in the season was interpreted as a low-risk strategy in birds that are less likely to face starvation. In the same species, the ratio of false alarms to real attacks decreased with the number of real attacks and on rainy days as both are expected to increase perceived predation risk. The effect of group size was not investigated in this system. One difficulty with an analysis of false alarms in foraging animals is that the foregone returns from interrupting foraging are not negligible. Therefore, factors affecting the value of foraging are likely to influence the choice to respond or not to potential threats making it more difficult to compare false alarms across conditions. A further difficulty is that false alarms in a foraging context may be used deceptively by individuals to usurp resources from others or to increase access to food (Mller 1988; Kahlert 2006; Wheeler 2009). Here, I propose an analysis of false alarms in resting animals where the only cost of responding to a potential threat is the energetic cost of fleeing. Deceptive use of alarm signals is unlikely in animals that are not competing for any resources but simply resting. I examine false alarms in staging semipalmated sandpiper (Calidris pusilla) flocks roosting on the shore at high tide. In particular, I examined the hypothesis that false alarms should be more frequent in larger groups and under environmental conditions that increase perceived predation risk, such as low light levels. Sandpipers double their body mass during fall staging (Hicklin 1987). This large increase in body mass will undoubtedly increa (...truncated)