Quantifying human disturbance on antipredator behavior and flush initiation distance in yellow-bellied marmots

Human disturbance may differentially affect the behavior of wild animals and such behavioral perturbations may have fitness consequences. To understand the effects of specific types of human disturbance on antipredator behavior, a behavior whose performance enhances survival, we studied yellow-bellied marmots (Marmota flaviventris). We quantified both antipredator vigilance and the flight initiation distance of the marmots to an approaching human in six different colony sites where we also quantified the frequency and type of human visitation. We developed an analysis framework, using linear mixed models, and found that: (1) when the presence of motorized vehicles and bicycles was high, marmots increased the proportion of time spent vigilant (pseudo R-2 = 0.33 and 0.31 for motorized vehicles and bicycles, P < 0.05) and decreased the time spent foraging (pseudo R-2 = 0.29 and 0.23 for motorized vehicles and bicycles, P < 0.05), (2) there was no significant effect of the presence of pedestrians on the time allocated to vigilance and foraging (pseudo R-2 = 0.25 and 0.19, P > 0.05), (3) marmots decreased the flight initiation distance as disturbance of motorized vehicles (pseudo R-2 = 0.85) and pedestrians (pseudo R-2 = 0.84) increased (P < 0.05), and (4) when we considered bicycles as the disturbance, juveniles tolerated closer approaches than adults or yearlings (P < 0.001). Marmots thus responded to some human disturbance by adjusting time spent in foraging and shortening the tolerance distance. Since these behavioral responses could have significant implications for survival and reproduction, we should generally view human disturbance as something that can influence natural antipredator behavior. Importantly, based on an understanding of the differential effects of human activities on wildlife, reducing human disturbance should be taken into account for wildlife management. In addition, our approach will be useful to quantify differential effects of humans on wildlife and to enhance our ability to manage those impacts. (C) 2010 Elsevier B.V. All rights reserved.