Understanding and Predicting Population Response to Anthropogenic Disturbance: Current Approaches and Novel Opportunities

Abstract

Effective conservation of biodiversity depends on the successful management of wildlife populations and their habitats. Successful management, in turn, depends on our ability to understand and accurately forecast how populations and communities respond to human-induced changes in their environments. However, quantifying how these stressors impact population dynamics remains challenging. Another significant hurdle at this interface is determining which quantitative approach(es) are most appropriate given data types, constraints and the intended purpose. Here, we provide a cross-taxa overview of key methodological approaches (e.g., matrix population models) and model elements (e.g., energetics) that are currently used to model the effects of anthropogenic disturbance on wildlife populations. Specifically, we discuss how these modelling approaches differ in their key assumptions, in their structure and complexity, in the questions they are best poised to address and in their data requirements. Our intention is to help overcome some of the methodological biases that might persist across taxonomic specialisations, identify new opportunities to address existing modelling challenges and improve scientific understanding of the direct and indirect impacts of anthropogenic disturbance. We guide users through the identification of appropriate model configurations for different management purposes, while also suggesting key priorities for model development and integration.

FIGURE 1

Conceptual framework for assessing the individual‐, population‐ and community impacts of human disturbance and management actions on animal populations. Adapted and extended from (National Academies of Sciences, Engineering, and Medicine 2017).

FIGURE 2

Considerations for ecological processes or mechanisms that should be considered depending on how disturbances impact a population. These considerations are non‐sequential and often interrelated, as indicated by the dashed boxes. Coloured boxes describe the different ways a population could be impacted or indicate key requirements. Grey boxes provide brief descriptions of when an ecological process may be relevant to include in a model.

FIGURE 3

Decision tree for identifying appropriate modelling approaches depending on the primary research goal. Note, these approaches are limited to the modelling approaches discussed within this paper and the recommendations are under the assumption that data availability is not a primary limiting factor. Structured population models include matrix population models and integral projection models. For each approach, we indicate whether the approach is spatially‐explicit (green icon) or complex (pink icon), whether energetics processes can be incorporated (yellow icon) and whether is it always the case (solid outline on icon) or can be (no outline).

FIGURE 4

Overview of common or promising modelling integrations for disturbance ecology. Modelling approaches are coloured according to the ecological scale that disturbance impacts are usually modelled (i.e., geographical range–pink; individual‐level–blue; population‐level–purple). Descriptions of why each approach may be beneficial to integrate are attached to each modelling approach. Examples of integrations are indicated along the connecting arrows, with the type of integration indicated by the arrow head (arrow–the output of one model is used as an input for the other model; closed circle–both models are integrated within a single multimodal framework with feedback between components; closed square–integration type can be as for the arrow or closed circle depending on the specific modelling approach or research question).

FIGURE 5

Summary of the red fox (Vulpes vulpes) case‐study model comparisons for (A) rabies and (B) culling, highlighting the primary data types used, the spatial structure, key processes represented, model outcomes and the key unknowns when modelling red fox population dynamics. Each model is colour coded to indicate which elements are specific to a single model (coloured) or shared between both models (grey).

FIGURE 6

Schematic for the European mink (Mustela lutreola) case study, highlighting the (A) range contraction of European mink and concurrent expansion of invasive American mink (Neovison vison) in France, and (B) potential energetic impacts of key threats to the European mink, illustrated using dynamic energy budget (DEB) theory. Threats are colour coded to indicate how they affect the energy budget: Green‐via ingestion (e.g., reduced food intake), orange‐via somatic maintenance (e.g., increased energy expenditure) and purple‐reproduction (e.g., reproductive impairment). The white boxes over the mink refer to key energy stores used in DEB theory. Maturity and somatic maintenance are energetic costs.