Determination of the Effectiveness of Wildlife Barriers and Underpasses on U.S. Highway 441 across Paynes Prairie State Preserve, Alachua County, Florida

Final Report

October 1999

Introduction

Massive numbers of wildlife roadkills have occurred on the section of U.S. Highway 441 across Paynes Prairie State Preserve (Franz and Scudder, 1977; Smith, 1996). The high levels of wildlife mortality have adversely affected wildlife populations, created human safety concerns, and detracted from the beauty of the prairie. This research project was designed to determine the effectiveness of a wildlife barrier/underpass system.  This phase of the project was designed to determine the level of pre-construction wildlife mortality and to determine how many and what types of animals use the existing culverts.

Methods

Road surveys consisted of a researcher walking the entire 3.2 km (2 mile) length of the road during each sampling unit (sampling unit = 24 hours).  A 3-day sampling period was conducted each week.  This consisted of a researcher going out on day 1 and marking all dead animals found, and on days 2 and 3 recording all roadkills. Thus 104 randomized sampling units per year were sampled during the year-long study (2, 24-hr sampling units x 52 weeks).

The area surveyed consisted of the entire road surface (north and southbound lanes) and extended 3-4 m onto the grassy shoulders (1 pass on each side of the road). The median also was surveyed (1 pass) on foot, or mountain bicycle. The road was divided into 100-m sections, for a length of 3.2-km.  The first section was located on the north rim of the prairie at the first private driveway (see Fig.1).

Surveys were conducted at first light and all live and dead animals were recorded. Dead animals were marked with spray paint so that they were not counted more than once.  The paint used was both lead and toluene-free (Forestry Suppliers, Jackson, MS).  Locations of all animals also were recorded, e.g., north or southbound lane, right-of-way, median, and 100-m section.

Freshly killed, undamaged specimens were collected (GFC Permit # WS98348), preserved, and deposited in the Florida Museum of Natural History at the University of Florida (see Appendix A for a list of specimens collected).

Funnel traps were installed in the northern and southern box culverts to capture amphibians, reptiles, and small mammals. In the southern culvert, ten equally spaced hardware-cloth traps were placed along the sides. Four floating screen funnel traps were installed in the northern box culvert which was inundated during the first 9 months of the study.  Ten hardware cloth traps were set in the northern culvert once water levels had receded.  The two culverts in the center of the prairie were completely inundated throughout the study and no traps were placed there.

The trapping schedule coincided with road surveys. However, to assure adequate sampling, the trapping period included an additional 2 sampling units (48 hours- for a total of 5 trap-days per week).  All animals captured (excluding venomous snakes) were marked, measured, weighed, and released. Age and sex of animals was determined whenever possible.  A sand track station (1.8 m long by 1 m wide) was maintained in the center of the southern, dry culvert.  The track station was monitored during the trapping schedule. A TrailMaster Active Infrared monitor and camera (Goodson and Associates, Inc., Lenexa, KS) were installed at the center of the southern culvert to monitor animal use of culverts throughout the year.

Nighttime observations were conducted as a supplement to road surveys. The purpose of these surveys was to examine the proportion of animals that successfully cross the road versus the number attempting to cross the road.  Florida Department of Transportation conducted traffic counts at the north rim of the prairie for one 24 hr period per month for the course of the study.  Traffic data are presented in Appendix B.

Environmental data was recorded in two locations along the roadway.  Air and water temperature, relative humidity, and barometric pressure were measured with computer data loggers (Onset Computer Corporation, Bourne, MA).

Results

A total of 3,365 dead animals (1,333 frogs; 29 alligators; 374 turtles; 1 lizard; 1,291 snakes; 265 birds; and 72 mammals) were counted from August 19, 1998 through August 13, 1999 (Table 1). This total includes all dead animals observed during the study (3 days per week x 52 weeks). The green tree frog Hyla cinerea was the most commonly observed species (13.8% of total), followed by Florida banded water snake Nerodia fasciata (12.7%), southern leopard frog Rana utricularia (11.5%), and green water snake Nerodia floridana (6.2%). These five species represented more than 50% of the total number of animals killed.

The mean number of kills per 24-hour sampling period was 17.3 (sdev = 15.5; range = 0 - 79). Monthly means ranged from 8.0 through 44.8 and were highest in July and August 1999(Fig. 2). Most carcasses were located in the outside lanes (55.5%) or bicycle lanes (21.7%), followed by the inside lane (13.0%) (Table 2). Very few were found in the median (1.5%), grassy right-of-way (4.6%), and centerline (3.7%).  Most carcasses were located in sections 4-6, on the north rim of the prairie, but large numbers of kills also occurred in sections 19-28 (Fig. 3).

Twenty-eight species were documented using the existing north and south box culverts (Table 3). Funnel trap captures included 50 animals in 3,950 trap-days (1.0% capture success). Most captures (76%) occurred between March and August. Tracks of nine-banded armadillo (Dasypus novemcinctus), river otter (Lutra canadensis), opossum (Didelphis virginianus) and raccoon (Procyon lotor) often were observed in the dry southern culvert. These four species also were repeatedly photographed with the motion sensor camera. Three small mammals, rice rat (Oryzomys palustris), hispid cotton rat (Sigmodon hispidus), and round-tailed muskrat (Neofiber alleni), used the dry culvert.

Although nighttime surveys were conducted, animals were not observed in sufficient numbers to estimate the success rate of animals attempting to cross the road.  The first nighttime survey was conducted on September 21, 1998 following 35 mm of rainfall.  Two eastern cottonmouths (Agkistrodon piscivorus) were observed on the right-of-way; however, no movement across the road was observed over a two-hour period (from 2200 – 2400 hrs). Two additional night surveys were conducted on 3 and 20 February 1999 (from 2200 – 2430 hrs).  On 3 February several southern leopard frogs were seen on the southbound traffic lanes and grassy right-of-way. No movement across the road was observed.  No live animals were observed during the 20 February night survey.  On 15 March and 17 April surveys were conducted from 2000-2300 hrs. No live animals were observed during these surveys.  Night surveys also were conducted on 12, 13, and 14 June (from 2100-2330 hrs). Several freshly killed eastern mud snakes (Farancia abacura), pig frogs (Rana grylio), southern leopard frogs, and eastern cottonmouths were observed. Live animals encountered included two pig frogs, a leopard frog, and one eastern cottonmouth. The eastern cottonmouth attempted to enter the northbound outside lane, but retreated after several vehicles nearly ran over it.

During daylight hours, live animals found on the right-of-way included eastern cottonmouth, southern leopard frog and American alligator (Alligator mississippiensis); however, these animals were not observed attempting to cross the road. A red fox (Vulpes vulpes), Florida box turtle (Terrapene carolina), and common snapping turtle (Chelydra serpentina) were observed successfully crossing the road.  A round-tailed muskrat was seen attempting to cross the road at 0700 hrs in May. The muskrat darted across both northbound lanes to the median before returning to the northbound right-of-way.