Elk Adjusted to Crop-Field Scare Devices Faster Than Black-Tailed Deer

IFarm Fields

Elk and Deer Lost Alertness Over 45 Days

Rutherford and coauthors tested the devices from August 4 to September 17 of 2023 in 15 crop fields near Duncan in the Cowichan Valley on Vancouver Island and published their work in Ecology and Evolution in 2025.

Our rule is to use a motion-triggered device on a food plot or crop field only in the weeks of greatest damage risk, use shouting clips and lights, review the videos weekly, and swap clips or shift the unit when animals remain through the playback. This rule comes from three results in the trial. Shouting caused the most fleeing, lights increased it, and alertness waned over 45 days of continued exposure. That rule went untested in the trial because the devices stayed in the same spots with the same 24 clips for all 45 days.

IIPicking Sounds

Shouts Outperformed Talking, Dogs, Predators and Birds

Every unit at a field edge combined a trail camera with a speaker and two LED light strips and ran one of 24 clips for 6 seconds. Clips came in sets of six for human voices (three talking and three shouting), aggressive dogs, wolf howls and cougar screams, and bird calls. A salt block brought animals into the camera's view.

Filtering left 680 events divided into 368 involving elk and 312 involving deer. With all sounds included, elk fled in 19% of events and deer in 43%.

Shouts made elk flee 36% of the time and beat every other sound, while playbacks of talking, dogs, predators and birds ran 13 to 17.5% with no differences among those sounds. Deer fled in response to 64% of shouts, 46% of talking and 45% of predator calls. Those shares are raw values. With bird calls as the comparison, modeled odds of deer flight were 4.4 times higher with shouting and 2.24 times higher with predator calls and both 95% confidence intervals stayed above 1. Talking and dog barks had intervals that overlapped 1, leaving neither sound with a clear advantage over bird calls for deer.

The authors interpret the low elk response to talking as a sign of long familiarity with human voices in a valley of farms, traffic and tourists. The authors suggest a shout may sound more threatening because it may go with chasing or thrown objects in conflict areas.

IIIAdding Lights

Lights Increased Flight Odds Without Changing Alert Time

Each site changed from sound alone to sound with lights or the reverse after 3 weeks, with eight sites beginning with lights and seven beginning without. Elk fled in a raw 26% of events with lights and 13% without. The authors' models put flight odds with lights at 4.7 times the odds without for elk and 1.8 times for deer while also accounting for sound, group size and site. Those values are odds ratios, which means the 4.7 does not say that 4.7 times as many elk fled.

Flight was more likely in bigger groups with odds ratios of 1.51 for elk and 1.45 for deer per standard deviation of group size and did not differ by age, sex or time of day.

The study scored an animal as alert when it was vigilant, running or out of frame. Lights did not change how long elk or deer stayed alert during or after a playback. Elk averaged 13% of the time alert before a playback, 67% during and 69% afterward and deer rose from 18% before to 81% during and 84% afterward.

IVWaning Alertness

Predicted Elk Alertness Dropped From 86.4% to 55.0% in 45 Days

The predicted chance of an elk being alert during and after a playback dropped from 86.4% at day 1 to 55.0% at day 45. For deer, that predicted chance dropped from 84.8% to 62.9%. Those changes mean relative declines of 36.3% for elk and 25.8% for deer.

Flight patterns separated the species. Elk became less likely to flee as the trial progressed (odds ratio 0.55 per standard deviation of trial day with a 95% confidence interval of 0.39 to 0.78). Deer had no clear drop in flight (odds ratio 0.94 with an interval overlapping 1).

Against bird calls, elk flight faded less over the weeks with predator calls (odds ratio 3.73) and dog barks (2.87). Shouting and talking showed the same direction for elk flight, though the two intervals overlapped 1. Lights may have kept flight from fading as quickly in both species because fleeing was somewhat more common late in the trial with the lights on. The authors describe this as weak evidence, with light intervals overlapping 1 for both species (odds ratios of 1.66 for elk and 1.53 for deer).

For deer, a faster decline in alertness was associated with more houses within 500 meters, while a slower decline was associated with greater distance from the highway. No clear pattern by site showed up in elk.

The authors also say managers may need to switch sounds or use devices with added reinforcement. Our read is that most landowners already have that reinforcement in a fence or lawful hunting pressure, though the trial tested neither. FDR's food plot design guide explains how to match acreage to browsing, and FDR's native browse guide covers why a plot adds to the browse surrounding it.

VStudy Boundaries

Crop Damage Went Unmeasured and a Full Season Went Untested

Because the trial tracked behavior at the device without measuring crop damage, it cannot tell whether animals ate less after fleeing. A salt block brought animals within camera range, while animals coming to a crop may respond differently. The authors say 15 sites and 10 stimulus combinations gave little statistical power for sound comparisons and the trial covered only 45 days in late summer with animals that were unmarked, so individuals could not be compared. The authors think repeated exposure allowed by the 1-minute gap between playbacks likely fed habituation and that the absence of any real consequence may have added to it. An archery-only elk season began September 1 midway through the trial, and the authors say it may have increased the likelihood that elk ignored the recorded threats. The trial studied black-tailed deer on Vancouver Island. It did not test whitetails.