Christopher T. Rota

Predicting Species Distributions from Samples Collected along Roadsides

Predictive models of species distributions are typically developed with data collected along roads. Roadside sampling may provide a biased (nonrandom) sample; however, it is currently unknown whether roadside sampling limits the accuracy of predictions generated by species distribution models. We tested whether roadside sampling affects the accuracy of predictions generated by species distribution models by using a prospective sampling strategy designed specifically to address this issue. We built models from roadside data and validated model predictions at paired locations on unpaved roads and 200 m away from roads (off road), spatially and temporally independent from the data used for model building. We predicted species distributions of 15 bird species on the basis of point-count data from a landbird monitoring program in Montana and Idaho (U.S.A.). We used hierarchical occupancy models to account for imperfect detection. We expected predictions of species distributions derived from roadside-sampling data would be less accurate when validated with data from off-road sampling than when it was validated with data from roadside sampling and that model accuracy would be differentially affected by whether species were generalists, associated with edges, or associated with interior forest. Model performance measures (kappa, area under the curve of a receiver operating characteristic plot, and true skill statistic) did not differ between model predictions of roadside and off-road distributions of species. Furthermore, performance measures did not differ among edge, generalist, and interior species, despite a difference in vegetation structure along roadsides and off road and that 2 of the 15 species were more likely to occur along roadsides. If the range of environmental gradients is surveyed in roadside-sampling efforts, our results suggest that surveys along unpaved roads can be a valuable, unbiased source of information for species distribution models.

The role of wildfire, prescribed fire, and mountain pine beetle infestations on the population dynamics of black-backed woodpeckers in the black hills, South Dakota.

Wildfire and mountain pine beetle infestations are naturally occurring disturbances in western North American forests. Black-backed woodpeckers (Picoides arcticus) are emblematic of the role these disturbances play in creating wildlife habitat, since they are strongly associated with recently-killed forests. However, management practices aimed at reducing the economic impact of natural disturbances can result in habitat loss for this species. Although black-backed woodpeckers occupy habitats created by wildfire, prescribed fire, and mountain pine beetle infestations, the relative value of these habitats remains unknown. We studied habitat-specific adult and juvenile survival probabilities and reproductive rates between April 2008 and August 2012 in the Black Hills, South Dakota. We estimated habitat-specific adult and juvenile survival probability with Bayesian multi-state models and habitat-specific reproductive success with Bayesian nest survival models. We calculated asymptotic population growth rates from estimated demographic rates with matrix projection models. Adult and juvenile survival and nest success were highest in habitat created by summer wildfire, intermediate in MPB infestations, and lowest in habitat created by fall prescribed fire. Mean posterior distributions of population growth rates indicated growing populations in habitat created by summer wildfire and declining populations in fall prescribed fire and mountain pine beetle infestations. Our finding that population growth rates were positive only in habitat created by summer wildfire underscores the need to maintain early post-wildfire habitat across the landscape. The lower growth rates in fall prescribed fire and MPB infestations may be attributed to differences in predator communities and food resources relative to summer wildfire.

A two-species occupancy model accommodating simultaneous spatial and interspecific dependence

Occupancy models are popular for estimating the probability a site is occupied by a species of interest when detection is imperfect. Occupancy models have been extended to account for interacting species and spatial dependence but cannot presently allow both factors to act simultaneously. We propose a two-species occupancy model that accommodates both interspecific and spatial dependence. We use a point-referenced multivariate hierarchical spatial model to account for both spatial and interspecific dependence. We model spatial random effects with predictive process models and use probit regression to improve efficiency of posterior sampling. We model occupancy probabilities of red fox (Vulpes vulpes) and coyote (Canis latrans) with camera trap data collected from six mid-Atlantic states in the eastern United States. We fit four models comprising a fully factorial combination of spatial and interspecific dependence to two-thirds of camera trapping sites and validated models with the remaining data. Red fox and coyotes each exhibited spatial dependence at distances > 0.8 and 0.4 km, respectively, and exhibited geographic variation in interspecific dependence. Consequently, predictions from the model assuming simultaneous spatial and interspecific dependence best matched test data observations. This application highlights the utility of simultaneously accounting for spatial and interspecific dependence.

A multispecies occupancy model for two or more interacting species

Summary Species occurrence is influenced by environmental conditions and the presence of other species. Current approaches for multispecies occupancy modelling are practically limited to two interacting species and often require the assumption of asymmetric interactions. We propose a multispecies occupancy model that can accommodate two or more interacting species. We generalize the single-species occupancy model to two or more interacting species by assuming the latent occupancy state is a multivariate Bernoulli random variable. We propose modelling the probability of each potential latent occupancy state with both a multinomial logit and a multinomial probit model and present details of a Gibbs sampler for the latter. As an example, we model co-occurrence probabilities of bobcat (Lynx rufus), coyote (Canis latrans), grey fox (Urocyon cinereoargenteus) and red fox (Vulpes vulpes) as a function of human disturbance variables throughout 6 Mid-Atlantic states in the eastern United States. We found evidence for pairwise interactions among most species, and the probability of some pairs of species occupying the same site varied along environmental gradients; for example, occupancy probabilities of coyote and grey fox were independent at sites with little human disturbance, but these two species were more likely to occur together at sites with high human disturbance. Ecological communities are composed of multiple interacting species. Our proposed method improves our ability to draw inference from such communities by permitting modelling of detection/non-detection data from an arbitrary number of species, without assuming asymmetric interactions. Additionally, our proposed method permits modelling the probability two or more species occur together as a function of environmental variables. These advancements represent an important improvement in our ability to draw community-level inference from multiple interacting species that are subject to imperfect detection.

Seeing the Forest through the Trees: Considering Roost-Site Selection at Multiple Spatial Scales

Conservation of bat species is one of the most daunting wildlife conservation challenges in North America, requiring detailed knowledge about their ecology to guide conservation efforts. Outside of the hibernating season, bats in temperate forest environments spend their diurnal time in day-roosts. In addition to simple shelter, summer roost availability is as critical as maternity sites and maintaining social group contact. To date, a major focus of bat conservation has concentrated on conserving individual roost sites, with comparatively less focus on the role that broader habitat conditions contribute towards roost-site selection. We evaluated roost-site selection by a northern population of federally-endangered Indiana bats (Myotis sodalis) at Fort Drum Military Installation in New York, USA at three different spatial scales: landscape, forest stand, and individual tree level. During 2007–2011, we radiotracked 33 Indiana bats (10 males, 23 females) and located 348 roosting events in 116 unique roost trees. At the landscape scale, bat roost-site selection was positively associated with northern mixed forest, increased slope, and greater distance from human development. At the stand scale, we observed subtle differences in roost site selection based on sex and season, but roost selection was generally positively associated with larger stands with a higher basal area, larger tree diameter, and a greater sugar maple (Acer saccharum) component. We observed no distinct trends of roosts being near high-quality foraging areas of water and forest edges. At the tree scale, roosts were typically in American elm (Ulmus americana) or sugar maple of large diameter (>30 cm) of moderate decay with loose bark. Collectively, our results highlight the importance of considering day roost needs simultaneously across multiple spatial scales. Size and decay class of individual roosts are key ecological attributes for the Indiana bat, however, larger-scale stand structural components that are products of past and current land use interacting with environmental aspects such as landform also are important factors influencing roost-tree selection patterns.

Apparent foraging success reflects habitat quality in an irruptive species, the Black-backed Woodpecker

ABSTRACT Dramatic fluctuations in food resources are a key feature of many habitats, and many species have evolved a movement strategy to exploit food resources that are unpredictable in space and time. The availability of food resources may be a particularly strong determinant of habitat quality for irruptive bird species. We studied the apparent foraging success of Black-backed Woodpeckers (Picoides arcticus), an irruptive species that responds opportunistically to pulsed food resources in burned forests and mountain pine beetle (MPB) infestations. Prior investigations revealed that the highest population growth rates of Black-backed Woodpeckers occurred in habitat created by summer wildfire, with intermediate population growth rates in MPB infestations, and the lowest population growth rates in habitats created by prescribed fire in fall. We tested whether apparent foraging success was associated with known habitat quality in order to assess the potential for food availability to regulate population growth. We counted the number of successfully captured wood-boring beetle larvae and “small” prey on each tree that a Black-backed Woodpecker used for foraging and modeled these counts as a function of habitat, tree diameter, number of years postfire, and tree disturbance category. Total apparent foraging success (the sum of observed captures of wood-boring beetle larvae and small prey per tree) did not vary across habitats, but woodpeckers foraging in habitats created by summer wildfire were expected to capture 2.2 and 2.0 times more wood-boring beetles than woodpeckers foraging in habitats created by fall prescribed fire and MPB infestations, respectively. These results suggest that the availability of food resources may contribute to population regulation in this irruptive species. Furthermore, population growth in irruptive species may be highly sensitive to the availability of preferred food resources. Forests recently burned by summer wildfires provide relatively abundant food resources for Black-backed Woodpeckers and represent high-quality habitat for this species of conservation concern.

A re‐evaluation of a case–control model with contaminated controls for resource selection studies

1. A common sampling design in resource selection studies involves measuring resource attributes at sample units used by an animal and at sample units considered available for use. Few models can estimate the absolute probability of using a sample unit from such data, but such approaches are generally preferred over statistical methods that estimate a relative probability of use. 2. The case-control model that allows for contaminated controls, proposed by Lancaster & Imbens (1996) and Lele (2009), can estimate the absolute probability of using a sample unit from use-availability data. However, numerous misconceptions have likely prevented the widespread application of this model to resource selection studies. We address common misconceptions regarding the case-control model with contaminated controls and demonstrate its ability to estimate the absolute probability of use, prevalence and parameters associated with categorical covariates from use-availability data. 3. We fit the case-control model with contaminated controls to simulated data with varying prevalence (defined as the average probability of use across all sample units) and sample sizes (n1 = 500 used and na = 500 available samples; n1 = 1000 used and na = 1000 available samples). We then applied this model to estimate the probability Ozark hellbenders (Cryptobranchus alleganiensis bishopi) would use a location within a stream as a function of covariates. 4. The case-control model with contaminated controls provided unbiased estimates of all parameters at N = 2000 sample size simulation scenarios, particularly at low prevalence. However, this model produced increasingly variable maximum likelihood estimates of parameters as prevalence increased, particularly at N = 1000 sample size scenarios. We thus recommend at least 500-1000 used samples when fitting the case-control model with contaminated controls to use-availability data. Our application to hellbender data revealed selection for locations with coarse substrate that are close to potential sources of cover. 5. This study unites a disparate literature, addresses and clarifies many commonly held misconceptions and demonstrates that the case-control model with contaminated controls is a viable alternative for estimating the absolute probability of use from use-availability data.

Characteristics of successful puma kill sites of elk in the Black Hills, South Dakota

Elk Cervus canadensis nelsoni in the Black Hills, South Dakota, have been declining since 2006 and there is concern by resource managers and hunters that puma Puma concolor predation may be contributing to declining herds. We evaluated characteristics at sites where puma successfully killed elk in the Black Hills of South Dakota. We evaluated characteristics at coarse (79-ha plots) and fine (0.2-ha plot) scales across the landscape. Our primary objective was to obtain a better understanding of vegetation and terrain characteristics that may have facilitated greater susceptibility of elk to predation by puma. We evaluated effects of road density, terrain heterogeneity, probability of elk use, and vegetation variables at 62 puma kill sites of elk and 186 random sites to identify key landscape attributes where elk were killed by puma. Elk were killed by puma in high use areas. Elk were also killed in areas that had greater amounts of edge and intermediate ruggedness at the coarse scale. Further, elk were killed in areas with greater small tree density and woody debris at the fine scale. High germination rates of ponderosa pine trees are unique to the Black Hills and provide dense patches of cover for puma. We hypothesize that cover from small trees and woody debris provided conditions where puma could stalk elk in areas with optimal security cover for elk. We suggest managers implement vegetation management practices that reduce small tree density and woody debris in areas with greater density of meadow—forest edge if they are interested in potentially diminishing hiding cover for puma in elk high use areas.

Mammal communities are larger and more diverse in moderately developed areas

Developed areas are thought to have low species diversity, low animal abundance, few native predators, and thus low resilience and ecological function. Working with citizen scientist volunteers to survey mammals at 1427 sites across two development gradients (wild-rural-exurban-suburban-urban) and four plot types (large forests, small forest fragments, open areas and residential yards) in the eastern US, we show that developed areas actually had significantly higher or statistically similar mammalian occupancy, relative abundance, richness and diversity compared to wild areas. However, although some animals can thrive in suburbia, conservation of wild areas and preservation of green space within cities are needed to protect sensitive species and to give all species the chance to adapt and persist in the Anthropocene.

Effects of an Extreme Flood Event on Federally Endangered Diamond Darter Abundances

Abstract Extreme flood events can substantially affect riverine systems, modifying in-stream habitat and influencing fish assemblages and densities. Rare species are especially vulnerable to these disturbance events because of their small population size and often reduced phenotypic heterogeneity. In June 2016 the lower Elk River in West Virginia experienced severe flooding, resulting in a peak discharge that exceeded the 0.005 annual exceedance probability (>200 y flood) in the main stem. We obtained pre-flood and post-flood population count data and estimated abundances for one cohort of the federally endangered Diamond Darter (Crystallaria cincotta) at 15 sites. While both the total count data and total estimated abundance decreased following the flood, our analyses did not indicate the extreme flood event strongly impacted Diamond Darter abundance. This indicates individuals are able to withstand high velocities and resist displacement or mortality. In addition site-level abundances were estimated at three sentinel sites during 2015 and 2016 using a multinomial N-mixture model that accounted for variation in detectability resulting from water temperature. Mean estimated abundance varied among the three sites and between the 2 y. Our results suggest there is substantial variation in year-class strength between the two cohorts we sampled. It is suggested that survey efforts at established sentinel sites be continued on an annual basis in order to help determine factors influencing year-class strength.