Bradley J. Udell

Wind turbines are a major source of bat fatalities in the United States, yet the magnitude and species-specific patterns of these fatalities have remained poorly quantified because monitoring has been site-specific, unstandardized, and not designed for landscape-level synthesis. In the most comprehensive analysis to date, we synthesized 499 postconstruction monitoring reports and developed a multispecies Bayesian hierarchical framework with an adapted N-mixture model to pool information across sites and estimate total fatalities for ten bat species across the conterminous United States from 1981 to 2024. Our analysis predicts approximately 18.7 million bat fatalities over this period, with substantially higher species-specific totals than previously estimated. Our framework provides a landscape-level analysis that can predict fatalities for unmonitored locations and years, identify landscape-level drivers of fatalities, produce a multidecade hindcast of cumulative losses due to wind energy, and map where species-specific risk is highest. Although limitations and uncertainty remain, the magnitude of predicted fatalities may be comparable to total losses from white-nose syndrome, another major threat to North American bat populations. These threats affect largely different species, suggesting their impacts may compound, but updated population-level impact assessments could reveal potential cumulative effects. Our estimates of species-specific covariate relationships with rotor-swept area, ground clearance, land cover, and proximity to rivers, highlight opportunities to help reduce impacts through wind turbine design and siting. This work establishes a generalizable landscape-level framework for evaluating wildlife fatalities at wind energy facilities and highlights how coordinated monitoring, data sharing, and population-level impact assessments could further inform conservation efforts.

A fundamental task in ecological statistics is to estimate abundance and growth rate distributions from wildlife monitoring data to inform conservation management. Modeling time series of wildlife populations presents a number of challenges from both statistical and ecological perspectives, including discreteness; lack of replication; nonstationarity; and observation, demographic, and other phenomenological processes. Nonstationary dynamics are often exhibited by populations undergoing environmental stressors. Models must account for these characteristics to produce reliable estimates of abundance and trends, yet estimation can be challenging with unreplicated data. We propose nonstationary demographic state-space models using unreplicated counts for populations undergoing environmental stressors. A reduced growth rate model matches the complexity of the unreplicated count data, and a fecundity bound on growth rate distributions allows the separation of processes affecting growth rates like environmental stressors from those affecting abundance external to growth rates like migration. NDSSMs allow for the embedding of nonstationary model components, and we explore the use of changepoints, volatility clustering, and migration processes. We apply the proposed nonstationary models in case studies of herons affected by predator/competitor reestablishment and three bat species affected by a fungal pathogen causing white-nose syndrome. Nonstationary models outperform stationary models and generalized linear mixed effects models according to model scoring and visual inspection of predictions, and provide estimates more consistent with published values. Incorporating migration improves model fit universally, even with approximate one-way immigration, most likely because populations are extirpated, recolonized, and increase multiple-fold over the upper bound set by species fecundity. In addition, estimates of the timing and severity of the environmental stressor differed for models with migration. Including nonstationary and demographic components in a fecundity-bounded growth rate model improves inference and benefits interpretability of hyperparameters. In turn, this adjusts uncertainties in predictions of abundance and growth rates over time, providing the ingredients needed for informed conservation analysis and for directing future monitoring of at-risk species. Supplementary materials accompanying this paper appear online.

Data describing habitat suitability are crucial for implementing effective conservation planning but are often lacking at regional and continental scales. We address this gap for 4 bat species that are listed, proposed for listing, or under Endangered Species Act listing review by highlighting a framework for estimating summer roost suitability with a presence-background approach to aid development of conservation policy and management plans. The 4 species of concern are the Little Brown Bat (Myotis lucifugus), the Northern Long-eared Bat (M. septentrionalis), the Indiana Bat (M. sodalis), and the Tricolored Bat (Perimyotis subflavus). Our estimates of summer roost suitability were developed for the eastern United States at a fine spatial resolution (250 m pixels) suitable for conservation planning across multiple scales. Summer roost habitat suitability was higher in areas with higher tree canopy cover for each of these species, though subtle differences were observed between the species that often use buildings (e.g., Little Brown Bat), tree crevices, cavities, and elements of dead or dying trees (e.g., Northern Long-eared Bat and Indiana Bat), and foliage (e.g., Tricolored Bat). To this end, roost suitability was not identical among species, and each showed subtly different relationships to the environmental covariates considered here. We also use a novel approach, gradient surface metrics, to quantify differences in the spatial pattern of summer roost suitability among the 4 species and found that tricolored bats and northern long-eared bats showed the most homogeneous and spatially smooth habitat suitability surfaces. Estimates of summer roost suitability developed here were also used to identify areas of good summer habitat where our models showed the least uncertainty that may be beneficial for targeted conservation, such as limiting disturbance to potential roost habitat. We also identified areas where additional data would benefit future summer roost modeling efforts. This work provides a first step toward developing multistate inventories of summer roost habitat suitable for implementing effective conservation planning at multiple scales.

Monitoring populations is challenging for cryptic species with seasonal life cycles, where data from multiple field techniques are commonly collected and analyzed as multiple lines of evidence. Data integration can provide comprehensive inferences while improving accuracy, precision, and scope but faces challenges in modeling misaligned resolutions and observational uncertainties. We developed a multi-scale, integrated species distribution model (MS-iSDM) for North American bats to combine data across monitoring types and seasons using joint likelihood methods, observational models with false-negatives and false-positives, and seasonal migratory connectivity. We applied this model to 11 years of data for an imperiled bat species (tricolored bat, Perimyotis subflavus). Relative abundance and occupancy were linked with multi-scale predictors, revealing clear patterns of population declines, but with important differences in spatial trends (abundance: corresponded with white-nose syndrome impacts, occupancy: at the range periphery) and overall severity (abundance: -74.8%, 95% CRI: -79.7% to -69.3%; occupancy: -35.5%, 95% CRI: -41.1% to -30.2%). The asynchrony between occupancy trends and population impacts was explained as an emergent pattern of spatiotemporal variation in abundance in the integrated distribution model. Compared to multiple lines of evidence, the integrated model provided consensus-estimates, increased precision and spatiotemporal scope, and strengthened evidence of population declines.

Estimating the abundance of unmarked animal populations from acoustic data is challenging due to the inability to identify individuals and the need to adjust for observation biases including detectability (false negatives), species misclassification (false positives), and sampling exposure. Acoustic surveys conducted along mobile transects were designed to avoid counting individuals more than once, where raw counts are commonly treated as an index of abundance. More recently, false‐positive abundance models have been developed to estimate abundance while accounting for imperfect detection and misclassification. We adapted these methods to model summertime abundance and trends of three species of bats at multiple spatial scales using acoustic recordings collected along mobile transects by partners of the North American Bat Monitoring Program (NABat) from 2012 to 2020. This multiscale modeling spanned individual transect routes, larger NABat grid cells (10 km × 10 km), and across the entire extent of modeled species ranges. We estimated relationships between species abundances and a suite of abiotic and biotic predictors (landcover types, climatological variables, physiographic diversity, building density, and the impacts of white‐nose syndrome [WNS]) and found varying levels of support between species. We present clear evidence of substantial declines in populations of tricolored bats (Perimyotis subflavus) and little brown bats (Myotis lucifugus), declines that corresponded in space and time with the progression of WNS, a devastating disease of hibernating bats. In contrast, our analysis revealed that similar population‐wide declines probably have not occurred in big brown bats (Eptesicus fuscus), a species known to be less affected by WNS. This study provides the first abundance‐based species distribution predictions and population trends for bats in their summer ranges in North America. These models will probably be applicable to assessing wildlife populations in other monitoring programs where acoustic data are used or where false‐negative and false‐positive detections are present. Finally, our abundance framework (as a spatial point pattern process) can serve as a foundation from which more sophisticated integrated species distribution models that incorporate additional streams of monitoring data (e.g., stationary acoustics, captures) can be developed for North American bats.

Removal sampling data are the primary source of monitoring information for many populations (e.g., invasive species, fisheries). Population dynamics, temporary emigration, and imperfect detection are common sources of variation in monitoring data and are key parameters for informing management. We developed two open robust‐design removal models for simultaneously modeling population dynamics, temporary emigration, and imperfect detection: a random walk linear trend model (estimable without ancillary information), and a 2‐age class informed population model (InfoPM, closely related to integrated population models) that incorporated prior information for age‐structured vital rates and relative juvenile availability. We applied both models to multiyear, removal trapping time‐series of a large invasive lizard (Argentine black and white tegu,Salvator merianae) in three management areas of South Florida to evaluate the effectiveness of management programs. Although estimates of the two models were similar, the InfoPMs generally returned more precise estimates, partitioned dynamics into births, deaths, net migration, and provided a decision support tool to predict population dynamics under different effort scenarios while accounting for uncertainty. Trends in tegu superpopulation abundance estimates were increasing in two management areas despite generally high removal rates. However, tegu abundance appeared to decline in the Core management area, where trapping density was the highest and immigration the lowest. Finally, comparing abundance predictions of no‐removal scenarios to those estimated in each management area suggested significant population reductions due to management. These results suggest that local tegu population control via systematic trapping may be feasible with high enough trap density and limited immigration; and highlights the value of these trapping programs. We provided the first estimates of tegu abundance, capture probabilities, and population dynamics, which is critical for effective management. Furthermore, our models are applicable to a wide range of monitoring programs (e.g., carcass recovery or removal point‐counts).

Poaching is a global problem causing the decline of species worldwide. Optimizing the efficiency of ranger patrols to deter poaching activity at the lowest possible cost is crucial for protecting species with limited resources. We applied decision analysis and spatial optimization algorithms to allocate efforts of ranger patrols throughout a national park. Our objective was to mitigate poaching activity at or below management risk targets for the lowest monetary cost. We examined this trade‐off by constructing a Pareto efficiency frontier using integer linear programming. We used data from a ranger‐based monitoring program in Nyungwe National Park, Rwanda. Our measure of poaching risk is based on dynamic occupancy models that account for imperfect detection of poaching activities. We found that in order to achieve a 5% reduction in poaching risk, 622 ranger patrol events (each corresponding to patrolling 1‐km² sites) were needed within a year at a cost of US$49,760. In order to attain a 60% reduction in poaching risk, 15,560 patrol events were needed at a cost of US$1,244,800. We evaluated the trade‐off between patrol cost and poaching risk based on our model by constructing a Pareto efficiency frontier and park managers found the solution for a 50% risk reduction to be a practical trade‐off based on funding constraints (comparable to recent years) and the diminishing returns between risk mitigation and cost. This expected reduction in risk required 8,558 patrol events per year at a cost of US$684,640. Our results suggest that optimal solutions could increase efficiency compared to the actual effort allocations from 2006 to 2016 in Nyungwe National Park (e.g., risk reductions of ~30% under recent budgets compared to ~50% reduction in risk under the optimal strategy). The modeling framework in this study took into account imperfect detection of poaching risk as well as the directional and conditional nature of ranger patrol events given the spatial adjacency relationships of neighboring sites and access points. Our analyses can help to improve the efficiency of ranger patrols, and the modeling framework can be broadly applied to other spatial conservation planning problems with conditional, multilevel, site selection.

Selective logging is the primary cause of tropical forest degradation and is rapidly expanding worldwide. While the impacts of logging on species diversity and distributions are well understood, little is known about the effects of logging on animal behaviours central to individual fitness and population persistence. The song rate of breeding songbirds is a behavioural trait that is often positively associated with male density and used by conspecific females as an indicator of territory and male quality. Thus, contrasting logging‐induced adjustments in song rates of individual birds with population shifts may illuminate potential mechanisms underlying population distributions. We present a novel application of bioacoustic monitoring, integrating counts of individuals, songs and duets from single automated recording units (ARUs) with N‐mixture models, to estimate shifts in population parameters (occupancy, abundance) and singing behaviours (per‐capita song rates, per‐pair duet rates) of 32 Bornean songbird species with logging. We tested hypotheses on the relationships between adjustments in behavioural and population parameters with logging, and further tested the extent to which species traits predicted behavioural and population shifts. Adjustments to singing behaviour in 59 and 53% of species (57% of duetting species) were concordant with differences in occupancy and abundance respectively, such that species showing reduced populations with logging also produced fewer songs per‐capita, and vice versa. Species known to prefer undisturbed habitats and large‐bodied species showed the most negative effects of logging on singing behaviour and population distributions. Species known to exploit degraded habitats exhibited the opposite pattern. Subdued singing in logged forests by species of conservation concern suggests limited competition between territorial males in small populations and may also signal low‐quality territories. Synthesis and applications. We demonstrate that bioacoustic monitoring can be used to not only estimate important population parameters of occupancy and abundance across a diverse tropical songbird community, but also enables quantification of behaviours considered relevant to individual fitness, yet unobtainable with conventional methods (e.g. point counts). Bioacoustics provides a viable approach to reliable automated large‐scale monitoring of hyperdiverse tropical forest systems under logging operations and other land‐use pressures.

Climate change and urban growth impact habitats, species, and ecosystem services. To buffer against global change, an established adaptation strategy is designing protected areas to increase representation and complementarity of biodiversity features. Uncertainty regarding the scale and magnitude of landscape change complicates reserve planning and exposes decision makers to the risk of failing to meet conservation goals. Conservation planning tends to treat risk as an absolute measure, ignoring the context of the management problem and risk preferences of stakeholders. Application of risk management theory to conservation emphasizes the diversification of a portfolio of assets, with the goal of reducing the impact of system volatility on investment return. We use principles of Modern Portfolio Theory (MPT), which quantifies risk as the variance and correlation among assets, to formalize diversification as an explicit strategy for managing risk in climate‐driven reserve design. We extend MPT to specify a framework that evaluates multiple conservation objectives, allows decision makers to balance management benefits and risk when preferences are contested or unknown, and includes additional decision options such as parcel divestment when evaluating candidate reserve designs. We apply an efficient search algorithm that optimizes portfolio design for large conservation problems and a game theoretic approach to evaluate portfolio trade‐offs that satisfy decision makers with divergent benefit and risk tolerances, or when a single decision maker cannot resolve their own preferences. Evaluating several risk profiles for a case study in South Carolina, our results suggest that a reserve design may be somewhat robust to differences in risk attitude but that budgets will likely be important determinants of conservation planning strategies, particularly when divestment is considered a viable alternative. We identify a possible fiscal threshold where adequate resources allow protecting a sufficiently diverse portfolio of habitats such that the risk of failing to achieve conservation objectives is considerably lower. For a range of sea‐level rise projections, conversion of habitat to open water (14–180%) and wetland loss (1–7%) are unable to be compensated under the current protected network. In contrast, optimal reserve design outcomes are predicted to ameliorate expected losses relative to current and future habitat protected under the existing conservation estate.

Better understanding human–wildlife interactions and their links with management can help improve the design of wildlife protection zones. One example is the problem of wildlife collisions with vehicles or human‐built structures (e.g., power lines, wind farms). In fact, collisions between marine wildlife and watercraft are among the major threats faced by several endangered species of marine mammals. Natural resource managers are therefore interested in finding cost‐effective solutions to mitigate these threats. We combined abundance estimators with encounter rate theory to estimate relative lethal collision risk of the Florida manatee (Trichechus manatus latirostris) from watercraft. We first modelled seasonal abundance of watercraft and manatees using a Bayesian analysis of aerial survey count data. We then modelled relative lethal collision risk in space and across seasons. Finally, we applied decision analysis and Linear Integer Programming to determine the optimal design of speed zones in terms of relative risk to manatees and costs to waterway users. We used a Pareto efficient frontier approach to evaluate the performance of alternative zones, which included additional practical considerations (e.g., spatial aggregation of speed zones) in relation to the optimal zone configurations. Under the various relationships for probability of death given strike speed that we considered, the current speed zones reduced the relative lethal collision risk by an average of 51.5% to 70.0% compared to the scenario in which all speed regulations were removed (i.e., the no‐protection scenario). We identified optimal zones and near‐optimal zones with additional management considerations that improved upon the current zones in terms of cost or relative risk. Policy implications. Our analytical framework combines encounter rate theory and decision analysis to quantify the effectiveness of speed zones in protecting manatees while accounting for uncertainty. Our approach can be used to optimize the design of protection zones intended to reduce conflicts between human waterborne activity and marine mammals. This framework could be extended to address many other problems of human–wildlife interactions, such as the optimal placement of wind farms to minimize collisions with wildlife or the optimal allocation of ranger effort to mitigate poaching threats.