1 Introduction
As the extent and magnitude of human activity continues to expand (IPBES
2019), the urgency to understand how wildlife populations respond to
anthropogenic change is accelerating (Larson et al. 2016; Venteret al. 2016). This information is crucial for effective
management and conservation policies (Pimm et al. 2014).
Ecologists have long tried to understand and predict the impacts of
human disturbance on wildlife populations and communities through the
use of quantitative modelling approaches (Beissinger & Westphal 1998;
Getz & Haight 1989). However, stressors rarely occur in isolation, with
animal populations often exposed to multiple direct (e.g., harvesting;
Kays et al. 2017) and indirect stressors (e.g., habitat
fragmentation, Smith et al. 2019). Although it is relatively
straightforward to determine how individual stressors impact
populations, the complex ways that multiple stressors interact make it
challenging to predict their combined effect on a population (Paniwet al. 2021). As a result, there has been an increasing focus on
understanding the population-level effects, such as changes in
population dynamics, geographical distribution, and/or population
persistence, that can result not only from individual indirect
(non-lethal) stressors but from exposure to multiple stressors (e.g.,
Gosselin et al. 2015; Daversa et al. 2025; Pirottaet al. 2019; Galic et al. 2018).
Oftentimes, disturbances have indirect effects on populations, such as
through changes in food intake, which may lead to changes in energy
balance and/or body condition (e.g., Parker et al. 2009), or
through exposure to pathogens or pollutants, which may result in changes
in immune status (Charbonnel et al. 2008). These indirect effects
can compound, leading to impacts on vital rates (e.g., survival,
reproduction) that then shape population dynamics (Fig. 1, Box 1).
Explicit consideration of these indirect effects has been formalised for
marine mammal risk assessments (NAS 2017). Other efforts have advocated
for the inclusion of indirect effects within a standardised mechanistic
framework (Johnston et al. 2019). However, these are conceptual
frameworks and, as such, lack information on specific approaches used to
model population-level impacts of disturbance. These frameworks also
fall short when considering community dynamics or management actions
(but see Urban et al. 2022), which can play key roles in
predicting population persistence (Fig. 1, Box 1).
Interactions among stressors and discrepancies in spatial and temporal
scales of impact mean that correlative approaches are often too limited
to inform policy or management. Quantitative modelling approaches thus
play an important role in understanding and managing ecosystems by
clarifying the key mechanisms that might explain the behaviour of
ecological systems (Schmolke et al. 2010). The main advantage of
quantitative approaches to management decision making is their ability
to predict the magnitude of effect of alternative scenarios based on
underlying processes, which is rarely possible through empirical studies
alone (Skogen et al. 2024). Many quantitative approaches are
available for modelling the impacts of human activity on animal
populations, each of which have their own assumptions, caveats, and
advantages. The approaches differ in which processes are represented
(and how), their spatio-temporal scales, data requirements, and in which
questions they are best poised to answer. Inappropriate choices in
modelling approach or structure could compromise our ability to make
reliable predictions (Gerber 2006), creating greater uncertainty when
deciding on management strategies.
Despite the vast array of modelling approaches available, several key
challenges remain when trying to predict the population-level impacts of
human activity on wildlife: 1) how to get the most out of disparate data
streams collected at different spatio-temporal scales; 2) how to provide
scientifically informed management advice for populations when limited
empirical data are available, as is typically the case; 3) how to manage
uncertainty when uncertainty is ubiquitous; and 4) knowing which
approach(es) to use given the available data and the question at hand.
Further challenges arise when considering community dynamics, which is
often necessary to accurately predict the implications of human
disturbance and potential management strategies on ecological
communities (Buckley and Han 2014; Zavaleta et al. 2001). While
challenges 1-3 have received considerable attention (e.g., Nichols 2021;
Simmonds et al. 2024; Zipkin et al. 2019; Fletcher Jr.et al. 2019), guidance on modelling choices for assessing and
predicting population-level impacts of human disturbance has received
less attention (but see Accolla et al. 2021; Hunter-Ayad et
al. 2020; Thompson et al. 2021; Briscoe et al. 2019).
Here, we provide an overview of key approaches available to model human
impacts on animal populations. In doing so, we aim to provide resources
for new studies to identify suitable methods and help overcome taxonomic
or domain biases in model development. As part of this effort, we
highlight important considerations when deciding on a modelling approach
and model structure to model the direct and indirect impacts of human
disturbance on animal populations. We also use two case-studies on red
fox (Vulpes vulpes ) and European mink (Mustela lutreola )
to illustrate some of the considerations made during model development
and the strategies used to overcome the limitations of different
modelling approaches, including model integration and energetic
modelling. Further, we extend existing conceptual frameworks for
understanding the impacts of human disturbance by incorporating
community-level responses to multiple stressors as well as the pathways
by which management actions can influence a population (Fig. 1, Box 1).
The information presented here can be used to identify appropriate model
configurations for different research and management purposes, while
also suggesting key priorities for future model development and
integration.