Farmland on potential paths went with longer dispersal

STEP 1 · Birth range

Natal agriculture showed no detectable departure effect

The landscape was about 60% agriculture and about 24% forest. Both papers counted pasture and hay together with cultivated crops as agriculture. Corn and soybeans were the main crops, and some collared deer ranged into northern Iowa. The researchers found no evidence that agriculture on the natal range influenced dispersal or migration. They suggest agriculture was already common on those ranges and may have left too little contrast to detect a departure effect. The methods also describe substantial spread in natal agricultural and forest share, so that explanation stays their suggestion.

The logistic models used 124 subadults (65 females and 59 males). The top models also found no detectable effect from the share of potential paths that crossed a road, or from natal range size. A second model kept migrants in the sample and linked larger natal ranges and a lower share of potential paths that crossed a road with higher dispersal odds. In the null models, natal range size was related to the occurrence or detection of dispersal. Our read leaves both of those predictors unsettled.

Opening results name 142 subadults and then list 41 dispersing males, 17 dispersing females, 3 migrating males, 16 migrating females, 26 resident males and 49 resident females. Those six counts add to 152. The paper says sample size differed by analysis. Our read does not turn that list into shares.

STEP 2 · Odds by sex

Young bucks had higher dispersal odds than does

Among 124 deer, males were more likely to disperse than females. The odds ratio was 2.77 and the 95% interval was 1.16 to 6.76. In the second model males had 2.74 times the odds of dispersing rather than staying put. The 95% interval was 1.18 to 6.39. Their estimated odds of migrating were 0.19 times the female odds. The interval ran from 0.04 to 1.03 and includes 1, so the study cannot say whether young bucks and does migrated at different rates.

The opening list has 16 migrating females and 3 migrating males. The authors also cite earlier work putting young doe dispersal near 21% in this Minnesota population and 4% in Wisconsin. No males in this sample migrated in autumn, and the authors warn that an autumn dispersal difference by sex rests on few females.

STEP 3 · Trip length

Agriculture on potential paths went with longer trips

Distance models used 56 dispersers (17 females and 39 males). Autumn trips were shorter than spring trips, and trips ran longer where simulated potential paths held more agriculture. Discussion text gives a mean of 20 kilometers (12.4 miles) for this study.

Every movement class avoided agricultural land cover relative to what was available, and dispersal showed the strongest avoidance. Migration looked similar, but the authors call the short migration records very sensitive and ask for caution. Residents and deer still on a pre-shift range selected closer streams and avoided higher ground. Steps during dispersal and migration showed no clear stream or elevation response.

STEP 4 · Wisconsin spring

Juvenile males left more often from agricultural ranges

Gilbertson and colleagues screened 590 juvenile, yearling and adult deer from captures in Iowa, Grant and Dane counties from 2017 to 2020. Of those 590 deer, 111 dispersed (18.8%). Dispersal reached 64.2% of juvenile males and was uncommon in other ages and sexes. Only males dispersed in fall. The median distance was 5.77 kilometers (3.6 miles), with a range from 1.3 to 68.3 kilometers (0.8 to 42.4 miles). The Minnesota 20 kilometer figure is a mean and this Wisconsin figure is a median from a different detector and age mix, so the two distances stay separate.

No deer in that Wisconsin screen showed an annual shift consistent with migration, so the sample cannot yield a migration rate. About 41.0% of the habitat was deciduous forest and 4.2% was mixed or evergreen forest. About 19.9% was pasture or hay and 21.8% was cultivated crops. The Minnesota authors describe that Wisconsin cover as near 45% forest and 42% agriculture against about 24% forest and 60% agriculture in southeastern Minnesota.

In spring only, more agricultural land on a juvenile male pre-dispersal range went with higher odds of leaving. Fall models had no clear driver. Distances were longer in spring and longer where potential paths held more agriculture. More neighboring deer went with shorter distances, and the authors do not treat that neighbor count as local density. Agriculture on potential paths was no longer statistically significant in the model that included body weight. Adding proximity and weight dropped that sample from 84 to 67 deer, a 20.2% reduction. A non-significant term in the smaller model does not show a zero path effect.

Dispersing juvenile males avoided agriculture. Pre-dispersal and non-dispersal steps did not show that avoidance. Dispersing males stayed nearer rivers and streams, and non-dispersal steps selected streams as well.

Both studies tied more agriculture on potential paths to longer trips and recorded avoidance of agricultural cover during the move. Minnesota did not detect a natal-agriculture effect on departure. Wisconsin found higher spring odds for juvenile males on pre-dispersal ranges with more agriculture. Ages, detectors, landscapes and models differ, so those probability results stay separate.

STEP 5 · Same deer next year

A matched return changes one hypothesis

Both studies measured relative avoidance of agricultural cover during the move itself. That avoidance describes the transit, not the cover where a settled deer is later found. The Wisconsin authors suggest unharvested crops may act as cover in fall. Individual Minnesota results suggested avoidance of agricultural cover weakened in autumn. Wisconsin individual models trended toward weaker avoidance in fall, though that pattern was not apparent for dispersal steps recorded every hour. Cameras on one farm cannot show that every buck fawn leaves agricultural ground.

The observation that changes the call is the same reliably identified deer back on home cameras in the matching season a year later. For that animal the pattern fits a seasonal shift better than a finished one-way leave. The return does not reproduce the collar classes, and it cannot prove the modeled odds or a population rate. A deer that never returns may have dispersed, died or gone unseen. Hunting was the main cause of death in the Minnesota study area. Neither paper measured hunting success. Where a deer was located is not a forecast of a sighting.