Lethal and non-lethal effects of predators and temporal trends in activity
We summarized direct cause-specific mortalities of tree voles by recording when a dead tree vole was observed in the presence of a predator on the nest platform. We interpreted these events as the strongest causal evidence of direct mortality of tree voles by a specific nest predator. To assess the potential response of nest predators to increases in tree voles in year 1, we assessed trends in seasonal and multi-annual detections of species at monitored nest platforms. We presented trend results as a multi-annual graph of monthly occupancy (proportion of monitored nest platforms occupied in a given month) and interpreted seasonal and annual peaks in detections for each taxa.
To examine non-lethal effects of nest predators, in addition to the previous predator detection analysis of detection/non-detection data, we summarized an index of tree vole activity at nest platforms (count of # detections per week) 12-weeks before and 12-weeks after a predator was detected and compared this among our three nest predators. We excluded zeros (weeks in which no tree vole was detected at a given nest platform) to avoid zero-inflating our count data.
Predators can adjust their temporal activity patterns to coincide with those of their main prey (Forsman, Anthony, Meslow, & Zabel, 2004). To examine whether temporal activity of predators coincided with those of their prey we created two density plots: 1) diel patterns of tree voles, flying squirrels, owls, weasels, and digging birds, and 2) diel patterns depicting high and low likelihood of predation (for details on assignment of low versus high likelihood, see Appendix S1) of tree voles by weasels.