Biological Conservation 226 (2018) 238–246 Contents lists available at ScienceDirect Biological Conservation journal homepage: www.elsevier.com/locate/biocon Fished species uniformly reduced escape behaviors in response to protection T ⁎ O. Kennedy Rhoadesa,d, , Steve I. Lonhartb, John J. Stachowiczc a University of California, Davis Bodega Marine Laboratory, 2099 Westside Road, Bodega Bay, CA 94923, USA b Monterey Bay National Marine Sanctuary, National Oceanic and Atmospheric Administration, 110 McAllister Way, Santa Cruz, CA 95060, USA c Department of Evolution and Ecology, University of California, Davis, 4330 Storer Hall, Davis, CA 95616, USA d Smithsonian National Museum of Natural History, Smithsonian Marine Station, 701 Seaway Drive, Fort Pierce, Florida 34949, USA ARTICLE INFO ABSTRACT Keywords: Predation is a critical ecological process that alters the structure and functioning of ecosystems through density- Behavior mediated and trait-mediated effects on lower trophic levels. Although studies have focused on harvest-driven Protected areas reductions in abundances and sizes of targeted species, human harvest also alters species morphologies, life Human harvest histories, and behaviors by selection, plasticity, and shifts in species interactions. Restricting harvest can recover Ecosystem function the biomass of targeted species, but it is less clear how behavioral phenotypes recover, particularly relative to the Flight initiation distance impacts of potentially opposing pathways of human influence. We investigated the effects of protection on the Fishing behavioral traits of a marine fish assemblage, recording behavior of 1377 individual fishes of nine targeted kelp forest species across 16 California marine protected areas (MPAs) varying in age, protection level, and diver visitation. With long-term, full protection from harvest, all fish species exhibited shorter flight initiation distance (FID, or the distance at which an animal flees from an approaching threat) and longer time delays before fleeing, despite differences in trophic position, microhabitat use, and other ecological characteristics. These escape behaviors were amplified across new MPAs regardless of protection level, suggesting that recovery is slow and likely the result of differences in genetic or early-life experience among individuals in these long-lived species. Although the effects of full protection from harvest were partially offset by recovering populations of large piscivorous predators, the net effect of long-term, full protection on fish behavior was shorter FID. Additionally, all species had shorter FID at sites more frequently visited by divers, and this effect was greater in sites with long- term, full protection from fishing. To the extent that escape behavior is correlated with foraging behavior and predation rates, these results suggest that human-induced behavioral changes may affect ecosystem processes, even after abundances have recovered. If recovery of ecosystem functioning and services are the management goal, assessments should be broadened to include the recovery of functional traits (including behavior). 1. Introduction assessing the impacts of extraction and predation on lower trophic le- vels (Peckarsky et al., 2008). Human and natural predation impacts ecosystem processes through Humans have overexploited a wide range of ecologically important direct reductions of prey densities (Estes et al., 2011) and through in- consumers, often targeting adult individuals with desirable phenotypes, direct effects on prey traits including morphology, life history, and and altering the morphological, size distribution, and behaviors of wild behavior (Breviglieri et al., 2017; Tollrian and Harvell, 1999). Although populations (Allendorf and Hard, 2009). These unique patterns of ex- conservation ecology has traditionally focused on the effects of human- traction have important indirect consequences for prey populations driven declines in predator biomass on ecosystem function, the mag- (Darimont et al., 2015), communities (Madin et al., 2016), and eco- nitude and direction of trait-mediated effects may differ from those of systems (Estes et al., 2011). Recent studies indicate that human-induced direct effects (Creel and Christianson, 2008), and the strength of trait- changes in the behavioral composition of targeted species may alter mediated and indirect effects often exceeds that of density-mediated species interactions and the structure of entire ecosystems (Heithaus and direct effects of predation on ecosystems (Preisser et al., 2005; et al., 2008; Madin et al., 2010). Non-extractive human activities may Schmitz et al., 2004; Werner and Peacor, 2003). It is therefore im- also indirectly favor certain behavioral traits in wild populations, portant to consider trait-mediated and indirect pathways when leading animals to favor flight or defense over foraging (Frid and Dill, ⁎ Corresponding author. E-mail addresses: [email protected] (O. Kennedy Rhoades), [email protected] (S.I. Lonhart), [email protected] (J.J. Stachowicz). https://doi.org/10.1016/j.biocon.2018.06.030 Received 1 September 2017; Received in revised form 12 May 2018; Accepted 27 June 2018 0006-3207/ © 2018 Elsevier Ltd. All rights reserved. O. Kennedy Rhoades et al. Biological Conservation 226 (2018) 238–246 2002), affecting the proportion of bold, aggressive and active pheno- which consists of MPAs that vary in age and protection level, and also types, and altering individual fitness, population dynamics, species in- experience recreational SCUBA diving (Menzel et al., 2013). MPA teractions, and ecological functioning (Madin et al., 2016; Sih, 2013; monitoring efforts have thus far focused on the recovery of species Wong and Candolin, 2014). composition and the biomass of fish, invertebrates, and algae, high- Management strategies that regulate extractive and non-extractive lighting increases in the biomass of targeted species across southern recreational activities may likewise have direct and indirect effects on California MPAs (Caselle et al., 2015). The rate of recovery of fish es- animal behavior and ecological processes. Restoring the densities of cape behaviors across entire assemblages of harvested species has not single top predatory species via gear restrictions or catch limits pre- been assessed. Because these behaviors may affect predation rates on dictably strengthens behaviorally-mediated trophic cascades by en- herbivorous invertebrates, an important ecological process maintaining hancing anti-predator behavior and reducing foraging in prey (Estes kelp forests (Steneck et al., 2002; Tegner, 2000), a better understanding et al., 2011; Ripple et al., 2014; Wirsing and Ripple, 2011). Compara- of whether, how, and how quickly these behaviors respond to protec- tively, protected areas that limit or prohibit hunting, fishing, and/or tion will also provide key insight into the recovery of ecosystem func- other human-related disturbances for entire ecological communities tioning in protected areas. within their boundaries may remove the selective pressures and dis- To investigate the impacts of protected areas on the behavioral turbances that alter the behavior of a range of targeted species traits and ecological function of a marine assemblage, we conducted (Bergseth et al., 2016; Harmelin et al., 1995; Lowry and Suthers, 2004). field surveys of 1377 individual kelp forest fishes of nine different By restoring natural densities of strongly interacting species (e.g., pre- targeted species across sixteen sites within central California, which fall dators and prey), protected areas also indirectly impact species inter- within MPAs of varying age, protection level, and SCUBA diver visita- actions and the behavioral composition of various trophic levels tion level. We specifically asked how aspects of escape behavior (e.g., (Baskett and Barnett, 2015). Protected areas that also promote non- FID, time delay before flight) and habitat use varied among targeted extractive human activities such as ecotourism can alter habitat use and species and as a function of MPA age, protection level, and SCUBA diver foraging of resident wildlife (Arlinghaus et al., 2016; Geffroy et al., visitation. We hypothesized that species at different trophic levels and 2015). These behaviorally-mediated direct and indirect effects have with different habitat use may exhibit distinct direct and indirect be- unforeseen consequences for ecological processes such as predation, havioral responses to protection, with variable consequences for key competition, and herbivory in protected areas (Hebblewhite et al., ecological processes such as predation. 2005; Ripple and Beschta, 2006), which have yet to be thoroughly in- vestigated (Ford and Goheen, 2015). 2. Materials and methods Within marine ecosystems, entire assemblages of species are har- vested; yet the impacts of fishing on animal behavior have been well- 2.1. Study sites studied solely for single species. Even less is known about the impacts of protection from fishing on the behaviors of targeted assemblages We conducted surveys from June to September 2014 on shallow, (Bergseth et al., 2016). Field studies of the impacts of fishing and rocky reefs between the Breakwater in Monterey, CA, USA protection from fishing have focused on flight initiation distance (FID), (36.609414 N, −121.8924 W) and Weston Cove in Carmel Highlands or the distance at which an animal flees from a perceived threat CA, USA (36.51103 N, −121.94486 W) (Fig. 1; see additional in- (Ydenberg and Dill,