Distribution and habitat associations of reptiles, amphibians, and fishes on public lands of east-central Mississippi.

摘要

I surveyed reptile, amphibian, and fish communities and quantified habitat variables from 2000--2003 on Tombigbee National Forest and Noxubee National Wildlife Refuge to determine fauna-habitat associations surrounding first and second-order streams in mature, mixed pine-hardwood forests on public lands in east-central, Mississippi. Habitat models were calculated for reptile and amphibian species richness and abundance and for 7 reptile and 7 amphibian species. Percentage basal area pine, a variable used to indicate pine management intensity, influenced negatively reptile abundance, amphibian species richness, and abundance of broadhead skinks, northern fence lizards, two-lined salamanders, and Mississippi slimy salamanders.; Fish communities were sampled with a backpack electro-shocker. Wet stream depth, channel width, stream flow rate, and percentage canopy closure were the most influential stream variables affecting fish species richness and abundance. Percentage basal area pine influenced negatively fish abundance and bluehead chub abundance and influenced positively creek chub abundance.; Reptile abundance was about evenly distributed at 25 m intervals from 0 to 100 m from streams. Half, 50.7%, of all reptiles detected by area transect surveys were within 25 m of streams. In contrast, amphibian abundance was much greater near streams, and 83.2% of all amphibians detected by area transect surveys were within 25 m of streams. I recommend a streamside management zone (SMZ) width of at least 30 m as a buffer to protect reptile and amphibian populations during timber harvest.; Area transect surveys and pitfall trap arrays were used to survey herpetofauna. Area transects (50 x 6 m) were located 0, 25, 50, 75, and 100 m from streams, and I detected 25 reptile species (1,226 individuals) and 16 amphibian species (825 individuals) with area transect surveys. Pitfall traps (15 m length, 4 19-L buckets) were located 0, 50, and 100 m from streams, and I detected 10 reptile species (184 individuals) and 10 amphibian species (384 individuals) with pitfall traps. Pitfall trap data were biased severely by bucket depredation. Infrared-triggered cameras placed at pitfall trap arrays indicated that raccoons frequently removed captured animals from buckets during survey periods. The magnitude of the bias is unknown.