Temporal scaling in analysis of animal activity

Abstract

Activity is a fundamental aspect of animal behavior, but whether different factors influence activity at different temporal scales is poorly understood as ecologically‐relevant multi‐temporal‐scale (MTS) activity analysis frameworks have not been developed. The hierarchy of limiting factors hypothesis postulates that the scale of resource selection reflects its importance for escaping limiting factors, but the hypothesis has not been tested by studies that examine activity across multiple temporal scales. We hypothesized that a MTS analysis framework would reveal that different factors influence activity at different temporal scales. We studied American martens Martes americana fitted with sensors that measured activity during winter at high temporal resolution. Using a MTS analysis framework, we showed that different factors influenced marten activity across three different biologically based, hierarchically‐ordered temporal scales (day, bout, and decision). Mortality risk influenced marten activity at the broader scales examined (day and bout), but not at the finest scale (decision), which is consistent with the hierarchy of limiting factors hypothesis. We also found evidence of cross scale consistency and support of multiple hypotheses at bout and decision scales, suggesting some factors operate across multiple scales and trade‐offs occur for some factors (e.g. risk for forage). Prey activity and thermal conservation influenced activity at the finer scales examined, but not at the day‐scale, and competition influenced activity only at the bout scale. Marten activity was polyphasic and occurred across the 24‐h period, which is probably due to digestive and metabolic constraints. Results from this and other studies indicate that activity is sensitive to temporal scale for a wide range of species. MTS frameworks should lead to a more comprehensive understanding of how different mechanisms influence animal activity, and ultimately, biological fitness.