Predicting occupancy for pygmy rabbits in Wyoming: an independent evaluation of two species distribution models

Steve Germaine; Drew Ignizio; Doug Keinath; Holly Copeland

doi:10.3996/022014-jfwm-016

Home/Archive/USGS

USGS · 70117702

Predicting occupancy for pygmy rabbits in Wyoming: an independent evaluation of two species distribution models

Steve Germaine·Drew Ignizio·Doug Keinath·Holly Copeland

View original source ↗Export metadata ↓

Abstract

Species distribution models are an important component of natural-resource conservation planning efforts. Independent, external evaluation of their accuracy is important before they are used in management contexts. We evaluated the classification accuracy of two species distribution models designed to predict the distribution of pygmy rabbit Brachylagus idahoensis habitat in southwestern Wyoming, USA. The Nature Conservancy model was deductive and based on published information and expert opinion, whereas the Wyoming Natural Diversity Database model was statistically derived using historical observation data. We randomly selected 187 evaluation survey points throughout southwestern Wyoming in areas predicted to be habitat and areas predicted to be nonhabitat for each model. The Nature Conservancy model correctly classified 39 of 77 (50.6%) unoccupied evaluation plots and 65 of 88 (73.9%) occupied plots for an overall classification success of 63.3%. The Wyoming Natural Diversity Database model correctly classified 53 of 95 (55.8%) unoccupied plots and 59 of 88 (67.0%) occupied plots for an overall classification success of 61.2%. Based on 95% asymptotic confidence intervals, classification success of the two models did not differ. The models jointly classified 10.8% of the area as habitat and 47.4% of the area as nonhabitat, but were discordant in classifying the remaining 41.9% of the area. To evaluate how anthropogenic development affected model predictive success, we surveyed 120 additional plots among three density levels of gas-field road networks. Classification success declined sharply for both models as road-density level increased beyond 5 km of roads per km-squared area. Both models were more effective at predicting habitat than nonhabitat in relatively undeveloped areas, and neither was effective at accounting for the effects of gas-energy-development road networks. Resource managers who wish to know the amount of pygmy rabbit habitat present in an area or wanting to direct gas-drilling efforts away from pygmy rabbit habitat may want to consider both models in an ensemble manner, where more confidence is placed in mapped areas (i.e., pixels) for which both models agree than for areas where there is model disagreement.

Explore related subjects

Journal of Fish and Wildlife Management Article

Source-reported bounding extent: 41.07935114946899° to 45.120052841530516° latitude; -111.26953125° to -103.974609375° longitude. This indicates report coverage, not an exact sampling location. View area on OpenStreetMap.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Steve Germaine, Drew Ignizio, Doug Keinath, Holly Copeland. 2014-08-01. Predicting occupancy for pygmy rabbits in Wyoming: an independent evaluation of two species distribution models. https://doi.org/10.3996/022014-jfwm-016

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2025

Lead exposure in waterfowl before contoxic shot requirements: A nationwide study, 1983−1986

Before implementing nontoxic shot requirements for hunting waterfowl and American coots Fulica americana in the United States in 1991, the U.S. Fish and Wildlife Service monitored lead poisoning in waterfowl on federal and state wildlife hunting areas during 1983-1986. Federal and state collaborators collected gizzards and livers from 9,029 hunter-killed waterfowl (10 species of dabbling ducks Anatinae , 9 diving ducks Aythyinae , 5 geese Anserinae , and tundra swans Cygnus columbianus ) across the four flyways. At the U.S. Fish and Wildlife Service National Wildlife Health Center, Madison, Wisconsin, waterfowl gizzards were examined for ingested lead and nontoxic shot and livers were analyzed for lead concentrations. Diving ducks had the greatest frequency (8.7%) of one or more ingested lead shot, followed by dabbling ducks (5.5%) and geese (1.3%). No ingested shot were found in tundra swans. The frequency of elevated (≥ 2.0 mg/kg wet weight) liver lead concentrations was also greatest in diving ducks, followed by dabbling ducks and geese. Within each species group, the frequency of elevated liver lead concentrations was greater than ingested lead shot, an indication that lead shot ingestion alone underrepresents lead exposure. Thus, lead in the liver may remain elevated after the erosion and excretion of lead pellets from the gizzard. Our results provide historical baseline data and summarize a nationwide study of lead exposure, using both ingested lead shot and liver lead concentrations, in waterfowl in the United States before the implementation of nontoxic shot regulations in 1991. These data can be compared with previous studies of lead exposure in waterfowl, as well as current and future assessments to evaluate the success of nontoxic shot regulations nationwide and specifically within previously sampled waterfowl management areas.

Journal of Fish and Wildlife Management↗

USGS2025

Exploration of a piscicide delivery method for invasive Black Carp

A piscicide delivery method was designed to selectively target Black Carp Mylopharyngodon piceus, an invasive fish species in North America, which possesses pharyngeal teeth adapted for crushing mollusk prey. Many North American mollusks exist in small, fragmented populations susceptible to local extirpation. A Black Carp-selective toxic bait deployed properly could be used to protect those populations. Results represent initial efforts to construct an effective single-dose species-selective toxic bait. We prepared a bait by attaching a glass vial containing toxicant (antimycin A) to the exterior of a Corbicula fluminea clam valve. We designed the vial to break from the force of the fish’s pharyngeal teeth when attempting to crush and consume the clam. We tested suitable vial sizes for encapsulating piscicide and two attachment materials, an aquarium epoxy and ultraviolet light-cured attachment material. The aquarium epoxy was rigid and reinforced vials. The ultraviolet light-cured attachment material was softer, resulting in more vials broken at 3- and 24-h post feeding. In a second trial, toxic baits consisting of antimycin A piscicide were administered in pure ethanol and acetone carriers to Black Carp in assembled baits with vials attached to live clam valves at three concentrations (40.0 mg/mL in ethanol, 170.8 mg/mL and 341.5 mg/mL in acetone). We selected aquarium epoxy for assembled baits based on the bond between the epoxy and glass vial, which allowed a greater surface area to be exposed and broken. While Black Carp successfully broke vials containing piscicide, no treatment caused carp mortality. The delivery method was unsuccessful as a single dose antimycin A piscicide bait for Black Carp. Additional considerations for this approach are provided.

Journal of Fish and Wildlife Management↗

USGS2025

Spatial variation of eDNA detection across an invasion gradient for invasive species monitoring programs

Spatial and temporal distribution data provide critical information for invasive species management. For example, distribution data can help managers with early detections and guiding other response actions. Environmental DNA (eDNA)-based sampling exists as one tool for monitoring invasive species. As part of bigheaded carp Hypophthalmichthys spp. monitoring efforts in the Illinois River, USA, we compared eDNA-based sampling results at multiple habitats across an invasion gradient in 2015. Greater densities of carp occurred downriver in the Illinois River and lower densities occurred upriver. We sampled from five locations along this gradient and from three habitat types (backwater, main channel, and shoreline) within each location. We sampled each location in April and June. A priori , we hypothesized that more eDNA detections would occur downriver, where higher densities of carp occur; that more eDNA detections would occur in backwater habitats compared to areas of the river with more fish movement; and that more eDNA detections would occur in April, because bigheaded carps are thought to use our sampling areas more during the spring. We compared the proportion of samples positive across this gradient, the habitat type, and the two sampling time periods. The most downriver location had the highest proportion of samples with eDNA detections, the backwater habitats had the highest proportion of samples with eDNA detections, and April had more positive detections than June. Our results highlight the importance of sampling across multiple habitat types and across time to gain a clear understanding of distribution when using eDNA-based sampling. Thus, being cognizant of the interactions between seasonal habitat use and eDNA-based detections is important for managers who rely upon eDNA-based monitoring.

Illinois↗