Letters, (2020) 23: 1522–1536 doi: 10.1111/ele.13571 REVIEWS AND SYNTHESES Ecological impacts of -induced behaviour change

Abstract Margaret W. Wilson,1* A growing body of literature has documented myriad effects of human activities on animal beha- April D. Ridlon,2 viour, yet the ultimate ecological consequences of these behavioural shifts remain largely uninves- Kaitlyn M. Gaynor,2 tigated. While it is understood that, in the absence of , variation in animal behaviour can Steven D. Gaines,1 have cascading effects on species interactions, community structure and ecosystem function, we Adrian C. Stier3 and know little about whether the type or magnitude of human-induced behavioural shifts translate Benjamin S. Halpern1,2 into detectable ecological change. Here we synthesise empirical literature and theory to create a novel framework for examining the range of behaviourally mediated pathways through which 1Bren School of Environmental human activities may affect different ecosystem functions. We highlight the few empirical studies Science & Management, University that show the potential realisation of some of these pathways, but also identify numerous factors of California, Santa Barbara, CA that can dampen or prevent ultimate ecosystem consequences. Without a deeper understanding of 93106, USA these pathways, we risk wasting valuable resources on mitigating behavioural effects with little 2 National Center for Ecological ecological relevance, or conversely mismanaging situations in which behavioural effects do drive Analysis and Synthesis, University of ecosystem change. The framework presented here can be used to anticipate the nature and likeli- California, Santa Barbara, CA hood of ecological outcomes and prioritise management among widespread human-induced beha- 93101, USA vioural shifts, while also suggesting key priorities for future research linking humans, animal 3Department of Ecology, Evolution, behaviour and ecology. and , University of California Santa Barbara, Santa Bar- bara, CA 93106, USA Keywords Animal behaviour, behavioural ecology, behavioural effects, ecosystem management, human impacts, human–wildlife interactions. *Correspondence: E-mail: mwwilso- [email protected] Ecology Letters (2020) 23: 1522–1536

limiting our ability to guide and prioritise management INTRODUCTION efforts. Without an enhanced understanding of the ecological As human activities continue to expand in magnitude, number consequences of human-induced behavioural effects, we risk and extent (Venter et al. 2016; Watson et al. 2016; Halpern both overlooking important drivers of ecological change that et al. 2019), a growing body of literature has documented are not addressed through traditional management strategies, widespread human impacts on animal behaviour across aqua- and misallocating management resources to mitigating beha- tic and terrestrial ecosystems (Wong & Candolin 2015; Larson vioural impacts that ultimately have little ecological relevance. et al. 2016; Shannon et al. 2016; Gaynor et al. 2018; Tucker While recent frameworks and case studies have linked et al. 2018; Samia et al. 2019; Suraci et al. 2019). Animal numerical declines of animal populations to ecological conse- behaviour underlies many critical ecological functions, includ- quences (Estes et al. 2011; Dirzo et al. 2014), we lack a similar ing nutrient cycling, primary productivity, pathogen transfer understanding of the behaviourally mediated pathways and habitat provision (Gribben et al. 2009; Barber & Dinge- through which humans impact ecosystems. Here, we present a manse 2010; Palkovacs & Dalton 2012). By affecting both novel framework outlining the pathways through which interspecific and intraspecific interactions, behavioural trait human activities may modify ecosystems via changes in ani- variation can alter population and community dynamics (Bol- mal behaviour. We begin by categorising the mechanisms nick et al. 2011) and wildlife conservation outcomes (Festa- through which human activities affect animal behaviour, syn- Bianchet & Apollonio 2003; Blumstein & Fernandez-Juricic thesising a broad literature on human-induced behaviour 2010; Berger-tal & Saltz 2016), yet we know little about change that previous reviews have segregated by ecosystem whether the type or magnitude of human-induced behavioural [e.g. (Marzano & Dandy 2012)], behaviour [e.g. flight shifts translate into ecological change. While many beha- (Stankowich 2008)] or human disturbance [e.g. noise (Wil- vioural effect studies allude to the implications of their find- liams et al. 2015; Shannon et al. 2016)]. We then present ings for populations, communities and ecosystems, limited detailed pathways linking documented animal behaviour empirical and theoretical investigation as well as a lack of syn- changes to established or hypothesised ecosystem conse- thesis across existing literature spheres preclude us from quences. While our integrative framework illustrates the knowing where and to what degree these impacts occur, potential for human-impacted behaviours to affect population dynamics, community interactions and ecosystem functions, [The copyright line for this article was changed on 22 September 2020 after we identify numerous factors likely to dampen these various original online publication] pathways and discuss the relevance of these factors for

© 2020 John Wiley & Sons Ltd. Reviews And Syntheses Ecological impacts of altered animal behaviour 1523 anticipating and managing the ecological consequences of induced changes in population density, though these potential behaviour change. While much remains to be learned about links have not been documented empirically. the drivers of animal behaviour change, we highlight compar- atively large knowledge gaps around the actualised ecological Top-down effects impacts of many human-impacted animal behaviours that prevent us from drawing management recommendations from Humans can have top-down impacts on animal behaviour by many existing studies. As this body of literature continues to directly or indirectly altering how and where perceive grow, we advocate for an increase in empirical and mod- risk [i.e. ‘landscapes of fear’ (Laundre et al. 2010; Gaynor et al. elling studies that go beyond documenting behavioural 2019; Suraci et al. 2019)]. Animals adjust their behaviours when impacts to examine the potential for realised ecological impli- they perceive direct human presence as a threat, even in response cations of human-induced animal behaviour change. to benign activities such as hiking or boating (Larson et al. 2016). Humans play the ecological role of ‘super predators’ in many systems and can shift the behaviour of even the highest MECHANISMS FOR HUMAN-INDUCED ANIMAL trophic level species (Darimont et al. 2015), triggering fear BEHAVIOUR CHANGE effects that can differ from and exceed those of natural preda- Human activities are increasingly impacting the aquatic and tors (Proffitt et al. 2009; Ciuti et al. 2012; Clinchy et al. 2016). terrestrial environments in which wildlife persist. In addition Humans can also indirectly affect a prey’s perception of risk by to our growing population and rising urban and agricultural modifying the populations of their natural predators, either development, technological advances enable us to access and increasing or decreasing risk of . For example modify previously remote environments (Ramirez-Llodra exploitation and habitat conversion have led to global predator et al. 2011; Pertierra et al. 2017), and increased participation losses (Estes et al. 2011), whereas predator restoration pro- in outdoor recreation expands our anthropogenic footprint in grammes [e.g. Canis lupus (gray