Time Allocation

For three summer months (from 20 May to 30 August 2009), and between 7:00 and 16:00 h, we conducted 2-min focal animal observation on foraging marmots (Blumstein et al., 2004). Each marmot was observed three times when it was foraging (they forage intermittently in the morning and late afternoon). Each colony was observed once every 3 days. The behaviors we recorded included foraging (stand foraging and rear foraging), vigilance (stand look and rear look), locomotion (walk and run), other miscellaneous behaviors, and periods of time when animals were out-of-sight behind rocks or vegetation. While observing a marmot, we dictated observations into microcassette recorders and later scored them using JWatcher 1.0 (Blumstein and Daniel, 2007). Before the behavioral observations were undertaken, the observers were trained to identify each behavior with 100% accuracy, and then trained with JWatcher until intra-observer scoring reliability was greater than or equal to 0.95.

Quantifying Flight Initiation Distance

We quantified the flight initiation distance (FID), the distance from an approaching threat at which an animal flees (Ydenberg and Dill, 1986) and measured FID of marmots in each colony. FID measures were taken independently to focal foraging observation. To minimize habituation to frequent disturbance, repeated observations had an interval of 5 days. Each marmot was approached and FID was measured at least three times, each on a different day. We approached subjects at 0.5 m/s (observers were trained to 95% accuracy before starting data collection), and dropped flags to note when they fled. We then used a laser rangefinder (Yardagepro 400, Bushnell Performance Optics, USA) to measure key distances that included start distance (the distance from the observer to the subject when the observer started walking towards it), flight initiation distance, and distance to burrow (the distance between the subject and the burrow when it initially moved). We used a clinometer to measure the incline where the marmot was when we initiated the approach. We categorized the incline as 0-10 degrees, 10-30 degrees and 30 degrees. Before doing our approach experiment, all observers were trained to walk consistently at 0.5 m/s.

Human Disturbance Data

At the peak of the tourist season (late June to mid July), we quantified human disturbance, every other day, for a total of 12 days. From 7:00 to 18:00 h, we continuously recorded the presence of motorized vehicles, bicycles and persons that stayed or passed through or within 300m of each colony and calculated the frequency of occurrence. Human disturbances were recorded independently of the behavioral observations and collection of FID data. From these observations, we calculated a pedestrians, bicycles, and total vehicles disturbance index for each colony. The index was calculated correcting the frequency of occurrence of each disturbance for the distance from the disturbance to the edge of the colony. The index calculated for pedestrians followed a bimodal distribution, so we categorized the colonies as having low or high pedestrian pressure. Moreover, considering the distribution of the total vehicular index, each colony was categorized as having low, medium or high vehicular pressure.

Statistical Analysis

We fitted linear mixed models, calculated with the function lmer from the package lme4 from the software package R version 2.10.1 (R Development Core Team, 2009). We ensured that the residuals of all models approximated to a normal distribution by visually checking normal probability plots and by the Shapiro-Wilk test. The proportion of time vigilant and foraging were arcsine transformed, whereas FID was [x0.2] transformed prior to analysis. To test for the effect of human disturbance on the proportion of time allocated to different activities we included in the model, as fixed factors, the marmot’s age and sex, and the index of human disturbance (total vehicles and pedestrians as categorical variables and bicycles as a continuous variable, one at a time). We included the interaction between the index of human disturbance and age, and when not significant, removed the interaction from the final model (Engqvist, 2005). For the analysis of flight initiation distance we also included as explanatory variables the observer’s starting distance, the incline of the land where the marmot was when we initiated the approach, and the distance the marmot was to its nearest burrow. In all cases (time allocation and flight initiation distance) the identity of the focal animal was included as a random factor to control for repeated measures. Moreover, the colony they came from was also included as a random factor. The significance of model parameters was estimated by comparisons to a probability distribution obtained by 10,000 Markov Chain Monte Carlo simulations with the function pvals.fnc from package language R (for further details see Baayen et al., 2008). The global models were compared using Nagelkerke’s pseudo R2 (Nagelkerke, 1991).

Major Results

Data were collected from a total of 96 individuals on 309 occasions (50 females: 23 adults, nine yearlings and 18 juveniles; 46 males: 11 adults, 14 yearlings and 21 juveniles).

• Marmots increased the proportion of time vigilant when presence of vehicles was high.
• No difference was found between the proportion of time spent vigilant under low and medium vehicular pressure.
• Marmots increased the time spent vigilant due to motorized vehicles and bicycles, and decreased the time spent foraging.
• Juveniles spent significantly less time vigilant than yearlings and adults.
• Marmots tolerated closer approaches as any type of human disturbance increased.
• Marmots showed a stronger response to pedestrians than to any other type of disturbance.

Next step – explore Case Study #2 (mayflies).

Attribution

Many thanks to undergraduate RMBL researcher, Danny DeSantiago, for illustrating this section and summarizing the major results. Thanks also to Dr. Dan Blumstein for his review and suggestions.