Mary M. Conner

Methods to quantify variable importance: implications for the analysis of noisy ecological data.

Determining the importance of independent variables is of practical relevance to ecologists and managers concerned with allocating limited resources to the management of natural systems. Although techniques that identify explanatory variables having the largest influence on the response variable are needed to design management actions effectively, the use of various indices to evaluate variable importance is poorly understood. Using Monte Carlo simulations, we compared six different indices commonly used to evaluate variable importance; zero-order correlations, partial correlations, semipartial correlations, standardized regression coefficients, Akaike weights, and independent effects. We simulated four scenarios to evaluate the indices under progressively more complex circumstances that included correlation between explanatory variables, as well as a spurious variable that was correlated with other explanatory variables, but not with the dependent variable. No index performed perfectly under all circumstances, but partial correlations and Akaike weights performed poorly in all cases. Zero-order correlations was the only measure that detected the presence of a spurious variable, whereas only independent effects assigned overlap areas correctly once the spurious variable was removed. We therefore recommend using zero-order correlations to eliminate predictor variables with correlations near zero, followed by the use of independent effects to assign overlap areas and rank variable importance.

MOVEMENT PATTERNS AND SPATIAL EPIDEMIOLOGY OF A PRION DISEASE IN MULE DEER POPULATION UNITS

Spatial patterns of disease occurrence across a landscape are likely products of both the ecological processes giving rise to underlying epidemics and the physical pathways of disease spread. Spatially explicit epidemic models often rely on assumptions about system boundaries and processes for spread that may not faithfully represent true patterns of host or vector distribution and movements. As a foundation for future modeling and parameter estimation, we evaluated potential influences of distribution and movements of mule deer (Odocoileus hemionus) on the spatial epidemiology of chronic wasting disease (CWD) in north-central Colorado. We used cluster techniques to define mule deer population units based on location data, and then used these as the sampling unit for subsequent analyses. We found marked differences in prevalence between population units that appeared at least partially related to deer movements. Migration (mean migration rate = 44%) rather than dispersal movements (≤2% dispersal rate) app...

Infectious Disease in Cervids of North America

Over the past two decades there has been a steady increase in the study and management of wildlife diseases. This trend has been driven by the perception of an increase in emerging zoonotic diseases and the recognition that wildlife can be a critical factor for controlling infectious diseases in domestic animals. Cervids are of recent concern because, as a group, they present a number of unique challenges. Their close ecological and phylogenetic relationship to livestock species places them at risk for receiving infections from, and reinfecting livestock. In addition, cervids are an important resource; revenue from hunting and viewing contribute substantially to agency budgets and local economies. A comprehensive coverage of infectious diseases in cervids is well beyond the scope of this chapter. In North America alone there are a number of infectious diseases that can potentially impact cervid populations, but for most of these, management is not feasible or the diseases are only a potential or future concern. We focus this chapter on three diseases that are of major management concern and the center of most disease research for cervids in North America: bovine tuberculosis, chronic wasting disease, and brucellosis. We discuss the available data and recent advances in modeling and management of these diseases.

RECOLONIZING WOLVES AND MESOPREDATOR SUPPRESSION OF COYOTES: IMPACTS ON PRONGHORN POPULATION DYNAMICS

Food web theory predicts that the loss of large carnivores may contribute to elevated predation rates and, hence, declining prey populations, through the process of mesopredator release. However, opportunities to test predictions of the mesopredator release hypothesis are rare, and the extent to which changes in predation rates influence prey population dynamics may not be clear due to a lack of demographic information on the prey population of interest. We utilized spatial and seasonal heterogeneity in wolf distribution and abundance to evaluate whether mesopredator release of coyotes (Canis latrans), resulting from the extirpation of wolves (Canis lupus) throughout much of the United States, contributes to high rates of neonatal mortality in ungulates. To test this hypothesis, we contrasted causes of mortality and survival rates of pronghorn (Antilocapra americana) neonates captured at wolf-free and wolf-abundant sites in western Wyoming, USA, between 2002 and 2004. We then used these data to parameterize stochastic population models to heuristically assess the impact of wolves on pronghorn population dynamics due to changes in neonatal survival. Coyote predation was the primary cause of mortality at all sites, but mortality due to coyotes was 34% lower in areas utilized by wolves (P < 0.001). Based on simulation modeling, the realized population growth rate was 0.92 based on fawn survival in the absence of wolves, and 1.06 at sites utilized by wolves. Thus, wolf restoration is predicted to shift the trajectory of the pronghorn population from a declining to an increasing trend. Our results suggest that reintroductions of large carnivores may influence biodiversity through effects on prey populations mediated by mesopredator suppression. In addition, our approach, which combines empirical data on the population of interest with information from other data sources, demonstrates the utility of using simulation modeling to more fully evaluate ecological theories by moving beyond estimating changes in vital rates to analyses of population-level impacts.

Population Dynamics of Spotted Owls in the Sierra Nevada, California

Abstract The California spotted owl (Strix occidentalis occidentalis) is the only spotted owl subspecies not listed as threatened or endangered under the United States Endangered Species Act despite petitions to list it as threatened. We conducted a meta-analysis of population data for 4 populations in the southern Cascades and Sierra Nevada, California, USA, from 1990 to 2005 to assist a listing evaluation by the United States Fish and Wildlife Service. Our study areas (from N to S) were on the Lassen National Forest (LAS), Eldorado National Forest (ELD), Sierra National Forest (SIE), and Sequoia and Kings Canyon National Parks (SKC). These study areas represented a broad spectrum of habitat and management conditions in these mountain ranges. We estimated apparent survival probability, reproductive output, and rate of population change for spotted owls on individual study areas and for all study areas combined (meta-analysis) using model selection or model-averaging based on maximum-likelihood estimation. We followed a formal protocol to conduct this analysis that was similar to other spotted owl meta-analyses. Consistency of field and analytical methods among our studies reduced confounding methodological effects when evaluating results. We used 991 marked spotted owls in the analysis of apparent survival. Apparent survival probability was higher for adult than for subadult owls. There was little difference in apparent survival between male and female owls. Model-averaged mean estimates of apparent survival probability of adult owls varied from 0.811 ± 0.021 for females at LAS to 0.890 ± 0.016 for males at SKC. Apparent survival increased over time for owls of all age classes at LAS and SIE, for adults at ELD, and for second-year subadults and adults at SKC. The meta-analysis of apparent survival, which included only adult owls, confirmed an increasing trend in survival over time. Survival rates were higher for owls on SKC than on the other study areas. We analyzed data from 1,865 observations of reproductive outcomes for female spotted owls. The proportion of subadult females among all territorial females of known age ranged from 0.00 to 0.25 among study areas and years. The proportion of subadults among female spotted owls was negatively related to reproductive output (no. of young fledged/territorial F owl) for ELD and SIE. Eldorado study area and LAS showed an alternate-year trend in reproductive output, with higher output in even-numbered years. Mean annual reproductive output was 0.988 ± 0.154 for ELD, 0.624 ± 0.140 for LAS, 0.478 ± 0.106 for SIE, and 0.555 ± 0.110 for SKC. Eldorado Study Area exhibited a declining trend and the greatest variation in reproductive output over time, whereas SIE and SKC, which had the lowest reproductive output, had the lowest temporal variation. Meta-analysis confirmed that reproductive output varied among study areas. Reproductive output was highest for adults, followed by second-year subadults, and then by first-year subadults. We used 842 marked subadult and adult owls to estimate population rate of change. Modeling indicated that λt (λt is the finite rate of population change estimated using the reparameterized Jolly–Seber estimator [Pradel 1996]) was either stationary (LAS and SIE) or increasing after an initial decrease (ELD and SKC). Mean estimated λt for the 4 study areas was 1.007 (95% CI  =  0.952–1.066) for ELD; 0.973 (95% CI  =  0.946–1.001) for LAS; 0.992 (95% CI  =  0.966–1.018) for SIE; and 1.006 (95% CI  =  0.947–1.068) for SKC. The best meta-analysis model of population trend indicated that λ varied across time but was similar in trend among the study areas. Our estimates of realized population change (Δt; Franklin et al. 2004), which we estimated as the product 1 × λˆ3 × λˆ4 × …× λˆk−1, were based on estimates of λt from individual study areas and did not require estimating annual population size for each study area. Trends represented the proportion of the population size in the first year that remained in each subsequent year. Similar to λˆt on which they were based, these Δˆt showed evidence of decline over the study period for LAS and SIE. The best model indicated recruitment of male and female adult and subadults varied from 0.10 to 0.31 new territorial individuals at time t/number of territorial individuals at time t − 1 and similarly among areas. We also conducted a population viability analysis (PVA) based on results of our meta-analysis. This PVA was of limited utility for ELD and SKC study areas because 95% confidence intervals on the probability of decline or increase spanned the interval [0, 1] within 5–10 years. When we restricted inferences to 7 years, estimated probability of a >10% decline for SIE was 0.41 (95% CI  =  0.09–0.78); for LAS the probability was 0.64 (95% CI  =  0.27–0.94). In contrast, estimated probability of a >10% increase in 7 years for SIE was 0.23 (95% CI  =  0.01–0.55) and for LAS was 0.10 (95% CI  =  0.00–0.34). For comparisons, we simulated a PVA for a hypothetical population with mean λ  =  1.0 and the same temporal variation as observed in our owl populations. Our PVA suggested that both the SIE and LAS populations had higher probabilities of declining in a 7-year period than increasing but that it would be difficult to determine if a population was in a slight gradual decline. Our analysis and the repository of information on our 4 study populations provide a data-rich template for managers to monitor impacts of future management actions on the owl. Specifically, our data can be used to evaluate the effect of management strategies on spotted owls that are being implemented by the United States Forest Service to reduce the risk of wildfire in the Sierra Nevada ecosystem. Our information also provides baseline information for evaluating the status of the owl for potential listing as a threatened species by the United States Fish and Wildlife Service.

EFFECTS OF FALL HUNTING ON SURVIVAL OF MALE WILD TURKEYS IN VIRGINIA AND WEST VIRGINIA

Abstract The Virginia Department of Game and Inland Fisheries and West Virginia Division of Natural Resources conducted a band-recovery study on male eastern wild turkeys (Meleagris gallopavo silvestris) during 1989–1996. Our main objectives were to estimate survival and band-reporting rates and to determine whether longer fall hunting seasons resulted in lower male turkey survival. Length of fall turkey hunting season varied from zero to 9 weeks at 3 study areas, while spring hunting season was relatively constant at 4 or 5 weeks. We attached reward leg bands to 473 male wild turkeys. Effects of different fall seasons were evaluated using survival and band-reporting rates. We used program MARK to construct a series of models including hunting-season structure, age (juvenile or adult), year, and period (fall or winter–summer) effects to estimate survival and band-reporting rates and to evaluate the effects of length of fall hunting on survival and band-reporting rates. Annual survival rates in our 3 study areas (range = 0.24–0.27) were lower than most studies. Survival estimates were significantly (P < 0.05) lower in the winter–summer (range = 0.16–0.18) than in the fall (range = 0.58–0.62). We found little difference in band-recovery estimates between age classes in the fall, but adults had significantly higher band-recovery estimates in the winter–summer. Male wild turkey mean annual survival did not decrease as fall hunting season length increased. In contrast, band-recovery rates increased as the length of the fall season increased. If band-recovery estimates indexed hunting mortality, then hunting mortality increased as length of fall hunting season increased. Moreover, if band-recovery rates represented hunting mortality, then the constancy of survival estimates among areas with different lengths of fall hunting season, coupled with the pattern of band-recovery rates, suggest that fall hunting mortality is not additive for male wild turkeys in Virginia and West Virginia, USA.

Fertility Control in Free-Ranging Elk Using Gonadotropin-Releasing Hormone Agonist Leuprolide: Effects on Reproduction, Behavior, and Body Condition

Abstract Overabundant elk (Cervus elaphus) populations have become a significant problem in many areas of North America. This is particularly true for protected areas where high densities of elk can cause long-term ecological degradation. When lethal control is not acceptable in these environments, resource managers must often consider alternative methods for reducing the size of resident elk populations. A potential management alternative is controlling the fertility of female elk. A promising new approach to wildlife contraception involves the use of biodegradable implants containing the gonadotropin-releasing hormone (GnRH) agonist leuprolide. During fall 2002–spring 2004, we compared pregnancy rates, reproductive behavior, daily activity patterns, and body condition of 17 free-ranging female elk treated with a leuprolide formulation with 17 untreated females, in Rocky Mountain National Park, Colorado, USA. After treatment, the pregnancy rate of treated elk was 0%, whereas 79% of control elk became pregnant. The effects of treatment were reversed the subsequent year with the pregnancy rate of treated females 100% compared with 91% for controls. Reproductive behaviors were similar for treated and control elk during the breeding and postbreeding seasons; treated elk did not exhibit postrut reproductive behaviors. Moreover, we found no differences in daily activity patterns of experimental groups during the breeding or postbreeding seasons. Treated elk did not show improved body condition over pregnant females. Instead, treated females catabolized proportionately more body fat during winter after treatment and at a higher rate than pregnant control elk. However, this effect was reversed the next spring with no difference in body fat between treated and control elk. We conclude from this experiment that leuprolide, administered as a controlled release formulation, offers a safe and effective method of controlling fertility in free-ranging female elk. However, practical application is limited by treatment duration and the need to treat females before the breeding season.

Effects of Gonadotropin-Releasing Hormone Immunization on Reproductive Function and Behavior in Captive Female Rocky Mountain Elk (Cervus elaphus nelsoni)

ABSTRACT Fertility control is a potential method for managing overabundant wildlife populations; however, current technology is limited by duration of treatment efficacy and unacceptable side effects. The objective of this study was to determine the efficacy of a single immunization with gonadotropin-releasing hormone (GnRH) vaccine to suppress reproductive function in pregnant female elk and to evaluate potential behavioral and pathological side effects of treatment. Eighteen captive adult female elk were randomly allocated to one of two experimental groups. Ten females were administered a conjugated and adjuvanted GnRH vaccine intramuscularly, and eight elk received an adjuvant sham vaccine without conjugated GnRH. We compared success of existing pregnancy, neonatal survival, subsequent fertility, reproductive behavior rates, and side effects of treatment between January 2006 and January 2010. The GnRH vaccination did not affect existing pregnancy or calf survival during the year that it was applied; however, it reduced the proportion of pregnant females for 3 yr. Male precopulatory behavior rates exhibited toward GnRH-vaccinated females tended to be greater than those directed at sham-vaccinated females during the second half of the breeding season, when GnRH vaccinates continued to be proceptive. Strong immune and inflammatory responses, including robust GnRH antibody concentrations in GnRH vaccinates, and sterile pyogranulomatous injection site abscesses in both groups, were consistent with vaccination. In conclusion, this GnRH vaccine resulted in prolonged, albeit reversible, impairment of fertility, and is associated with extended reproductive behaviors and partial suppression of hypothalamic-pituitary-gonadal axis function in captive female elk.

DISTINGUISHING ZAPUS HUDSONIUS PREBLEI FROM ZAPUS PRINCEPS PRINCEPS BY USING REPEATED CRANIAL MEASUREMENTS

Abstract Prebles meadow jumping mouse (Zapus hudsonius preblei), a federally listed threatened subspecies, and the western jumping mouse (Zapus princeps princeps) typically occur parapatrically but in some areas may be syntopic. Field differentiation between the taxa is difficult so we investigated the use of cranial characteristics as a basis for identification. We developed a discriminant function conducted on the means of repeated measurements to distinguish between the taxa from an initial sample of 105 specimens (n = 71 Z. p. princeps and n = 34 Z. h. preblei). We found that measurement error can contribute significantly to erroneous reclassification of specimens when only a single measurement set is used. Use of only presence or absence of the anterior median toothfold of M3 is not a reliable method for distinguishing between the subspecies. We used the discriminant function to identify 8 of 16 specimens collected in southeastern Wyoming as Z. h. preblei.

An occupancy-based quantification of the highly imperiled status of desert fishes of the southwestern United States

Desert fishes are some of the most imperiled vertebrates worldwide due to their low economic worth and because they compete with humans for water. An ecological complex of fishes, 2 suckers (Catostomus latipinnis, Catostomus discobolus) and a chub (Gila robusta) (collectively managed as the so-called three species) are endemic to the U.S. Colorado River Basin, are affected by multiple stressors, and have allegedly declined dramatically. We built a series of occupancy models to determine relationships between trends in occupancy, local extinction, and local colonization rates, identify potential limiting factors, and evaluate the suitability of managing the 3 species collectively. For a historical period (1889-2011), top performing models (AICc) included a positive time trend in local extinction probability and a negative trend in local colonization probability. As flood frequency decreased post-development local extinction probability increased. By the end of the time series, 47% (95% CI 34-61) and 15% (95% CI 6-33) of sites remained occupied by the suckers and the chub, respectively, and models with the 2 species of sucker as one group and the chub as the other performed best. For a contemporary period (2001-2011), top performing (based on AICc ) models included peak annual discharge. As peak discharge increased, local extinction probability decreased and local colonization probability increased. For the contemporary period, results of models that split all 3 species into separate groups were similar to results of models that combined the 2 suckers but not the chub. Collectively, these results confirmed that declines in these fishes were strongly associated with water development and that relative to their historic distribution all 3 species have declined dramatically. Further, the chub was distinct in that it declined the most dramatically and therefore may need to be managed separately. Our modeling approach may be useful in other situations in which targeted data are sparse and conservation status and best management approach for multiple species are uncertain.