Optimal Monitoring for Wildlife Biologists

ROOM: HCCC, Room 25B
Using data to fit and evaluate statistical models that represent our understanding of ecological processes is fundamental to wildlife ecology and management. Yet, monitoring designs for collecting data and statistical models are typically developed independently of each other. When monitoring designs are developed independently of statistical models, monitoring effort can be inefficient and fail to capture essential spatial, temporal, or spatio-temporal variability of an ecological process. Formally linking monitoring designs with statistical modeling using an optimal monitoring design permits several advantages over traditional monitoring methods including: increased efficiency, more precise parameter estimation, and reduced prediction uncertainty, all of which are critical in wildlife research and management. Furthermore, when financial resources limit the effort that can be devoted to collecting data, classical design-based inference may result in estimates that are insufficiently precise for management or conservation. Optimal monitoring designs permit extraction of the most information from the data that can be affordably collected, helping to improve our understanding of the ecological system, despite limited or shrinking financial resources. In this symposium, we provide an overview and history of optimal monitoring as well as future horizons in wildlife ecology, and four case studies that employ novel and state-of-the-art methods for explicitly linking monitoring designs with statistical modeling to improve inference and prediction in wildlife ecology and management. We examine applications of optimal monitoring to some of the most common sampling designs including occupancy surveys, abundance surveys, genetic sampling, and species distribution sampling.

8:10AM Optimal Adaptive Monitoring: Past, Present, and Future
  Mevin B. Hooten
Researchers have been using sequential strategies in the design of experiments and observational studies for decades. However, many of these approaches have arisen implicitly as a result of personal trial and error based on outcomes of former data collection efforts and analyses. In clinical trials and operations research a formal theoretical basis has been developed for optimally changing treatments midstream. However, in more complicated observational settings where spatio-temporal ecological processes are continually evolving as the system is being monitored, adaptive design strategies for simpler experiments may not apply. In atmospheric and environmental science, adaptive monitoring methods have been developed to take advantage of the estimated spatial structure and/or spatio-temporal dynamics in the system being studied (i.e., spatio-temporal fluctuations in air quality). Only recently have these methods been developed and applied in explicitly ecological studies. In this presentation, I motivate the use of formal model-based adaptive monitoring strategies and set the stage for the presentations that follow. I provide historical examples of these approaches and revisit a specific example involving a fish and wildlife habitat restoration effort by Missouri Department of Conservation to optimally monitor an ecological community along the Missouri River.
8:30AM Optimal Sampling Design for Autonomous Recording Units and Traditional Point-Count Sampling
  Beth E. Ross; Jesse M. Wood
Dynamic survey designs based on adaptive sampling practices yield increased efficiency, potentially reducing resources needed to monitor wildlife populations. Additionally, new technologies to survey wildlife, such as autonomous recording units (ARUs), may further optimize resources if combined with traditional survey methods such as point-count surveys. While ARUs can be used to monitor changes in species across time, developing a monitoring program combining ARUs with point counts has yet to be explored. The goal for our project was to develop an optimal sampling framework based on an occupancy model using different data sources with variable observational quality. To achieve this goal, we developed an occupancy model based on ARU and point-count surveys conducted on avian species in the Piedmont region of South Carolina. The latent presence/absence state was modeled to include covariates related to forest management practices in the region (e.g., thinning and burning of loblolly pine), while detection probability incorporated covariates related to wind, time of day, and survey type (ARU or point count). We used the output of our occupancy model to evaluate different designs for sampling with ARUs and point counts within our study area. Our results indicated differing relationships with forest management practices for different bird species (e.g., Yellow-breasted Chat and Indigo Bunting), resulting in different optimal designs depending on the species. Given the interest from stakeholders in understanding responses of the bird community to forestry practices, we present further results for the optimal design of songbird diversity in the study area. Overall, our approach presents a method to optimally allocate resources for wildlife based on different survey methodologies and could be applied to other remote technologies such as camera traps or unmanned aerial vehicles.
8:50AM Integrating Auxiliary Data in Optimal Spatial Design for Species Distribution Modeling
  Brian Reich; Krishna Pacifici; Jon Stallings
  • Traditional surveys used to create species distribution maps and estimate ecological relationships are expensive and time consuming. Citizen science offers a way to collect a massive amount of data at negligible cost and has been shown to be a useful supplement to traditional analyses. However, there remains a need to conduct formal surveys to firmly establish ecological relationships and trends.
  • In this paper, we investigate the use of auxiliary (e.g. citizen science) data as a guide to designing more efficient ecological surveys. Our aim is to explore the use of opportunistic data to inform spatial survey design through a novel objective function that minimizes misclassificaton rate (i.e. false positives and false negatives) of the estimated occupancy maps. We use an initial occupancy estimate from auxiliary data as the prior in a Bayesian spatial occupancy model, and an efficient posterior approximation that accounts for spatial dependence, covariate effects, and imperfect detection in an exchange algorithm to search for the optimal set of sampling locations to minimize misclassification rate.
  • We examine the optimal design as a function of the detection rate and quality of the citizen‐science data, and compare this optimal design with several common ad hoc designs via an extensive simulation study. We then apply our method to eBird data for the brown‐headed nuthatch in the Southeast US.
  • We argue that planning a survey with the use of auxiliary data improves estimation accuracy and may significantly reduce the costs of sampling.
  • 9:10AM Latent Spatial Models and Sampling Design for Landscape Genetics
      Ephraim M. Hanks
    We propose a spatially-explicit approach for modeling genetic variation across space and illustrate how this approach can be used to optimize spatial prediction and sampling design for landscape genetic data. We propose a multinomial data model for categorical microsatellite allele data commonly used in landscape genetic studies, and introduce a latent spatial random effect to allow for spatial correlation between genetic observations. We illustrate how modern dimension reduction approaches to spatial statistics can allow for efficient computation in landscape genetic statistical models covering large spatial domains. We apply our approach to propose a retrospective spatial sampling design for greater sage-grouse (Centrocercus urophasianus) population genetics in the western United States
    9:30AM Monitoring Dynamic Spatio-Temporal Ecological Processes Optimally
      Perry J. Williams
    Population dynamics vary in space and time. Survey designs that ignore these dynamics may be inefficient and fail to capture essential spatio-temporal variability of an ecological process such as population spread. Alternatively, dynamic survey designs incorporate knowledge of ecological processes, the associated uncertainty in those processes, and can be optimized with respect to monitoring objectives. We present a cohesive framework for monitoring a spreading population that links statistical diffusion models (and the associated uncertainty in those models) with survey design and monitoring objectives to reduce model uncertainty. We applied the framework to design optimal surveys for sea otters in Glacier Bay and Katmai National Parks in Alaska. By explicitly linking statistical diffusion models and survey design, we are able to reduce model uncertainty associated with forecasting occupancy, abundance, and colonization dynamics compared to other potential random designs.

    Organizers: Perry Williams, Colorado State University, Fort Collins, CO
    Supported by: TWS Biometrics Working Group

    Location: Huntington Convention Center of Cleveland Date: October 9, 2018 Time: 8:10 am - 9:50 am