Atlanta Coyote Pup Dens Sat Near Cover and Off Heavy Pavement

STEP 1 · Pup Filter

Only holes with pups entered the sample

The study sat inside I-285 where Fink and colleagues put density at about 7,170 people per square mile (about 2768 people per km2). The same section reports greater Atlanta tree canopy at 47.9 percent and calls that the highest share of any large US city. That canopy figure is cited there to a 2014 paper we did not open. Our read is that a city with little brush may not repeat this den pattern.

Spring GPS clusters pointed crews to candidate holes. A hole stayed in the analysis only when pups were seen. Holes that might still have belonged to another mesocarnivore were left out. Pup age at the checks and the timing of those clusters put birth in this Georgia sample from mid-March to mid-April. The checks ran from April 4 through April 29.

STEP 2 · Core Ground

Every den sat in the ground each coyote used most outside denning season

Every coyote with a pup den in this sample denned inside its 50 percent core. Fink and colleagues define that core as the space where the animal spends half its time. It is not half the home range by area. Average core size was about 914 acres (3.7 km2) with a range of about 262 to 2,335 acres (1.06 to 9.45 km2).

Those cores were fit to GPS locations taken outside March through August. The paper treats that span as den preparation and pup rearing and marks the span as unpublished. Fall and winter locations framed where the spring den sat. Each den was compared only with other ground inside the same core. The model did not test the core against the rest of the home range.

One coyote had no locations in the window used for cores. For that animal the core was a circle of median core size centered on the den. That den sits inside the core because the circle was drawn on the den.

Our read is that repeated winter sightings on the same ground are a rough proxy for heavy use. They are not a GPS core. Available cover and uneven time spent looking will change what gets recorded.

STEP 3 · Cover And Pavement

Dens pulled toward mapped cover and off heavy pavement

Inside those cores the coyotes selected dens nearer mapped cover and avoided high pavement. Mapped cover was dense forest plus scrub and brush. Each cell on that grid was 98 ft (30 m) on a side. Pavement here is shorthand for the impervious share of each cell in the National Land Cover Database, a relative measure of development. The covariates were centered and scaled before the model ran. A coefficient on that scale is not a change per foot. Selection for nearer cover was significant at p = 0.012. Avoidance of high pavement was significant at p = 0.036.

Half of the 20 dens were within 33 ft (10 m) of mapped cover, including the 7 that sat inside the cover class. One den was more than 0.62 mi (1 km) from mapped cover. Those are distances to the mapped cover class and not a tape run to a shrub. The authors give roads, parking lots and large commercial complexes as discussion examples of high-pavement ground. Those examples are not a tally of the den pixels.

Average distance from a den to a building was 443 ft (135 m) with a range from 42 ft (12.93 m) to 1,659 ft (505.56 m). Distance to the nearest building was not significant in this model (p = 0.459). Most locations inside these cores already sat within 492 ft (150 m) of a building. Fink and colleagues suggest this closeness likely kept coyotes from physically avoiding buildings at the scales they measured. Failure to detect a difference is not evidence that the effect is zero. Fink and colleagues say building type, occupancy and surrounding cover could change risk at a finer scale than these layers measured.

STEP 4 · Hard Material

Almost half the dens used discarded man-made material

Eleven of the 20 dens (55 percent) were natural structures. Those were flat-ground burrows, hillside burrows and standing or fallen trunks. The paper gives the remaining share as 45 percent anthropogenic. That group included discarded concrete, an overturned boat and a tractor tire. Six of the anthropogenic dens included some concrete. A mixed burrow was classed as anthropogenic when 50 percent or more of the structure was man-made material. The deepest den measured was an abandoned road culvert at 82 ft (24.9 m). Every anthropogenic structure in the study was abandoned or discarded.

Fink and colleagues suggest that hard roots and concrete added strength and protection. They present that suggestion as a presumption and they did not test it. They read the man-made material as a way to shield pups alongside selection for cover and avoidance of heavy pavement. They do not read the material as tolerance of people nearby.

STEP 5 · Distant Watch

A far look can check our rule where you can already walk

Our rule keeps the watch off parks. A national recreation area borders the northwest side of the study area and the paper credits that edge with significant forest. The paper does not report a den inside the park.

Birth timing in this Georgia sample ran from mid-March to mid-April and the den checks ran from April 4 through April 29. Our timing for a look follows that spring stretch here only as an observation window. Showing up in that window does not promise a pup.

Watch the woods and brush edge from a distance and do not approach the hole. Note discarded concrete, tires, boats and culverts along that edge. Write down whether a nearby building is occupied and what cover surrounds it. Crews searched holes with a flashlight or a burrow camera. They removed pups when they could under a university animal-care protocol. Our version is a distant look for pups at an entrance with the hole left alone.

A coyote pup in view shows that the hole was in use while you watched. The sighting cannot name the coyote or stand in for the selection model. Our cover-first rule would weaken if the pup dens you confirm kept landing on mostly paved ground far from cover or beside occupied homes while nearby cover went unused. Fink and colleagues report they could not determine how den decisions affected fitness for these Atlanta coyotes. They name disease burden, maternal age and parental experience as factors they did not score.