Cold Nights and Long Routes Raised Montana Doe Migration Odds

STEP 1 · Three-month spread

Most starts fell in late September and early October

Anton, DeCesare and Peterson kept 102 autumn migrations from 68 females after screening. Cabinet-Fisher contributed 34, the Rocky Mountain Front 40 and Whitefish 28. The three populations are partially migratory. Some animals in those areas stay on one range all year, and the model uses the females that left. The mean start was 19 October. The standard deviation was 22 days and screened starts ran from 17 September to 29 December. The paper found no significant difference in mean start day among the three areas. That result falls short of showing the herds share one date.

The return to winter range covered a three-month span across archery, rifle and closed periods in the study years. The model-predicted probability of migration rose steeply in late September and early October and then leveled off in late November and December. A start counted as the first directed move off the summer home range toward winter range when the deer stayed gone and reached winter range.

STEP 2 · Colder nights

Lower nightly lows and wetter weeks raised the odds

After the model accounted for date, deer were more likely to leave when the daily minimum temperature fell (P = 0.044) and when weekly precipitation rose (P = 0.027). Daily minimum temperature was the top model in the temperature set. Weekly precipitation ranked first in the precipitation set. Seven precipitation and snow-water formulations fell inside their 2 AIC cutoff for a competitive model. The combined model kept daily minimum temperature and weekly precipitation. Snow-water measures did not enter that final model.

In northwest Montana the authors report that temperatures often drop below freezing before significant snow accumulates. They suggest animals may start long routes in association with that cold, ahead of deep snow on high summer range. Snow-water measures were competitive within the precipitation set but were not carried into the final model, and the authors note that freezing nights often come before significant snow here. The study does not show that these does waited for deep snow. An end-of-season greenness term stayed in the top model and was not statistically significant (P = 0.154). Failure to detect a significant forage effect does not show the effect is zero. The weather terms were scaled before fitting, so the paper gives no degree mark or rainfall total to watch.

Our proposed log is optional and local. When you glass one basin, record the overnight low at a station you can repeat. Record whether that week was wetter than the week before. Record whether the does you know are still up high. Across several colder or wetter spells and more than one fall, that log shows the departures you saw and the weather that came with them. It cannot recover the fitted multiplier, prove a cause, name a collared individual or stand in for all three herds. If the departures you record keep missing those colder and wetter spells across multiple falls, treat this Montana pattern as a poor fit for that basin. One doe that leaves on a warm dry week settles nothing, because the sample's starts ran for months and individual deer differed.

STEP 3 · Route length

Longer migrations were more likely to start sooner

Migration distance in the screened sample averaged 16.3 miles (26.3 km). The standard deviation was 7.1 miles (11.5 km) and the range was 2.4 to 36.2 miles (3.9 to 58.2 km). Longer distance raised the chance of starting (P < 0.001). In a separate comparison of route measures, linear distance outranked the number of days spent migrating.

The authors calculated migration distance for each deer. The end of a migration was the first GPS location inside the winter home range, so the distance is known only after the deer arrives. Use route length only when marked-animal or agency migration data already show that does from that basin make a long trip. Casual sightings cannot link a summer doe to a winter doe. Individual deer also differed in baseline timing. Body condition data tied to age and fat were not available, so the study cannot explain early departures by age or fat.

STEP 4 · Quieter days

Departure odds were higher when hunting activity was lower

The authors predicted that higher hunting activity would raise the chance of leaving. The result ran counter to that prediction. Lower daily hunting activity went with a higher chance of starting (P = 0.0005). Autumn migration timing peaked in late September and early October, in line with lower hunting activity before the study-period rifle season peaked. A polynomial time term fit better than a split of the calendar into archery season and rifle season.

During the study, harvest of adult female mule deer was closed in the Cabinet-Fisher and the Whitefish Range and limited on the Rocky Mountain Front. The activity measure was meant to capture general disturbance from the influx of hunters. Licensed mule deer hunting in most of these districts focused on antlered males, and the estimates also included elk and white-tailed deer effort. The authors said hunting activity likely varies with land ownership, road access and terrain at a finer scale than they could measure. Our read is an association for these females. It does not test harassment and it does not estimate a hunter's personal odds. Drop the weather cue for a basin when the does you watch across several falls leave as rifle-season activity rises while nightly lows and weekly precipitation stay flat. That local record still cannot replace the fitted association for these collared females. Every migration in the model came from a doe.

STEP 5 · District lines

Collared deer in each area spanned up to 9 districts

Collared deer in each study area together spanned up to 9 hunting districts in a year. Individuals averaged 2.1 districts a year in the Cabinet-Fisher, 2.8 on the Rocky Mountain Front and 2.0 in Whitefish. The individual range was 1 to 4 districts.

On the Rocky Mountain Front during the study, several deer summered in district 150 where only antlered harvest was permitted. Many of those deer migrated to district 425 where harvest of females was permitted. The authors present the case as an illustration from the study period. Longer distances to reach summer range in district 150 would generally mean earlier autumn starts. The authors raised the case as a management concern. The district 150 population was below objectives during the study, and female harvest on winter range could hamper its growth depending on migration timing.

Our proposed observation is to mark the summer basin you glass. Once the does you watch are gone, look on the winter ground you already know from past years and write down which district that ground sits in. Layovers of a day or two were common on these migrations, so an empty morning on that winter ground does not map the route. Unmarked does on winter ground cannot be matched to the does you glassed in summer, so this record gives the winter district, not the route or its distance. Resident deer that stay all year sit outside the migratory sample.