Flooded Arkansas Swamp Rabbits Shifted to Uplands and Levees

STEP 1 · Earlier Picture

Stump beds were older observations and Illinois shrinkage was untested

Zollner and colleagues wrote that swamp rabbits had been seen resting on stumps surrounded by water. They also cited a southern Illinois telemetry study that described much smaller ranges during floods. That cited work ran no statistical test because the samples were small. That smaller-range description is only what this Arkansas paper reports from the Illinois study.

Pellet surveys in Alabama and an Indiana pellet survey reported by personal communication had not found swamp rabbits using surrounding uplands. The Arkansas study set out to measure how an individual's home range changed with flooding in south-central Arkansas. The authors described the species as popular game in the southern part of its range and a conservation concern in the northern part.

STEP 2 · Flood Mark

Inundation meant most of Moro Bottoms was under standing water

The rabbits were tracked on Moro Bottoms Natural Area and adjacent private land near Fordyce in Cleveland County. The natural area was about 70 ha (about 173 acres) and 40 ha of it (about 99 acres) was mature bottomland hardwoods. Methods place the natural area under shared stewardship.

Second-growth forest of mixed upland pines and bottomland hardwoods ran for several miles to the north and east. Sweetgum dominated the bottomland overstory. Loblolly pine dominated the adjacent upland overstory. Those sentences describe the woods around the bottom. They do not identify the cover rabbits used on the uplands or the levees in a flood.

Zollner and colleagues counted a location as inundated only when a permanent gauge in the center of the study area was deeper than 39 cm (about 15 inches). That depth matched times when more than 90 percent of the natural area held at least 20 cm of standing water (about 8 inches). The gauge recorded depth every hour. Flooding occurred during both seasons they analyzed.

They classed inundation from the water record on each location across the year. The paper does not report whether the shift differed between seasons. Our read is to check high ground when standing water covers most of the bottom. A single wet slough is a different case. That gauge depth belonged to this site alone.

STEP 3 · Location Shift

Two rabbits used creek levees and four used adjacent uplands

Zollner and colleagues reported that two rabbits moved onto levees near the creek during inundation and four moved into adjacent uplands. Permutation tests showed that each rabbit's flood locations were distributed differently from its dry locations. A test of that kind asks whether two groups of points come from different distributions. It does not show that flood ground sat entirely apart from dry ground.

The paper gives no share of the flood fixes on high ground and no distance for the moves. Moro Bottoms sat beside second-growth forest for several miles to the north and east.

They called the work the first quantitative assessment of swamp rabbit response to flooding and the best evidence they had that swamp rabbits use adjacent uplands during flooding. That claim belongs to this 2000 paper. Pellet surveys at the same site also showed upland use. Pellet surveys in Alabama and Indiana cited here had not.

STEP 4 · Kernel Size

A nonsignificant size test leaves range growth unsettled

Separate 95 percent fixed-kernel ranges were calculated for inundated locations and for dry locations. A paired test did not find a significant difference in home-range size. The printed test was t = 1.56 with 4 degrees of freedom and P = 0.18. The paper does not explain 4 degrees of freedom for a six-rabbit comparison. The authors said the estimates tended to increase during inundation. The authors suggested that their tendency toward larger flood ranges, set against the Illinois report of restricted flood ranges, could reflect different topographic or landscape contexts. The missing significance leaves range size unsettled for these six rabbits.

STEP 5 · Edge Check

Fresh sign shows local use and cannot name a rabbit

Our proposed check compares fresh sign on the upland edge and the levee before water covers most of the timber, while the bottom is flooded, and after the water drops. Repeat the comparison on more than one flood. New sign that shows up on that high ground only while the bottom is under water supports checking the same ground on a later flood at that site. Sign already on the high ground during dry days does not show a flood shift. Pellets that were already weathered before the water rose do not count as a flood shift. If repeated floods bring no new sign to the high ground, the Arkansas pattern does not fit that bottom.

What that walk can establish is the spread of fresh rabbit pellets and tracks on the ground you checked. Unless the sign can be assigned to swamp rabbits, it may also reflect other rabbits using the uplands. It cannot identify individual rabbits. It cannot rebuild a home range or copy the radio result. It cannot tell you whether a hunt will move a rabbit.

The paper cannot settle a bottom with no adjacent upland and no levee. Night locations were not estimated as precisely as day locations. Five of the six tracked adults were females. The male was tracked only in spring and summer. Four rabbits were tracked in both seasons. The remaining rabbit was tracked only in fall and winter. Each rabbit was located once a day until death or transmitter failure.

The authors asked managers to weigh landscape context and flood history before treating a bottom as enough habitat on its own. They suggested refuge habitat may matter where prolonged heavy inundation is common. They connected small patches in an inhospitable landscape with the northern part of the range.