this article was originally published in the encyclopedia ......encyclopedia of animal behavior...

This article was originally published in the Encyclopedia of Animal Behavior published by Elsevier, and the attached copy is provided by Elsevier for the

author's benefit and for the benefit of the author's institution, for non-commercial research and educational use including without limitation use in

instruction at your institution, sending it to specific colleagues who you know, and providing a copy to your institution’s administrator.

All other uses, reproduction and distribution, including without limitation

commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution’s website or repository, are

prohibited. For exceptions, permission may be sought for such use through Elsevier's permissions site at:

http://www.elsevier.com/locate/permissionusematerial

Frid A. and Heithaus M.R. (2010) Conservation and Anti-Predator Behavior. In: Breed M.D. and Moore J., (eds.) Encyclopedia of Animal Behavior, volume 1,

pp. 366-376 Oxford: Academic Press.

© 2010 Elsevier Ltd. All rights reserved.

Author's personal copy

Conservation and Anti-Predator BehaviorA. Frid, Vancouver Aquarium, Vancouver, BC, CanadaM. R. Heithaus, Florida International University, Miami, FL, USA

ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions

Deciding how much to invest in antipredator behavior,rather than in foraging or other activities, is a fundamentalproblem faced by most animals. Behaviors that loweran animal’s probability of death by predation – vigilance,hiding, avoidance of risky sites, and others – carry the costof reduced access to resources for growth and repro-duction. Ecologists have approached this problem, usingevolutionary principles. They have developed preda-tion risk theory, often expressed mathematically, whichassumes that prey maximize fitness (e.g., lifetime repro-ductive success) by making behavioral decisions that opti-mize trade-offs between predator avoidance and resourceacquisition. Predation risk theory can be used to estimatethe ‘risk effects’ of predators on prey: the lost foragingopportunities and lower levels of growth and reproductionexperienced by prey investing in antipredator behavior(also known as nonlethal or nonconsumptive effects).Importantly, risk effects can be major components of tro-phic cascades: the indirect effects of top predators on thepopulation processes of plants and animal species at lowertrophic levels, as mediated by the density and foragingbehavior of intermediate consumers (i.e., herbivores andmesopredators, collectively termed mesoconsumers). Thus,risk effects can potentially affect community structure andecosystem function.

This article surveys how human influences can affectantipredator behavior, potentially altering the risk effectsand mortality rates experienced by prey and the indirecteffects of predators on ecosystems. For example, humanconsumption of space and resources (e.g., fishing, forestry,agriculture, urbanization, etc.) and climate change maycreate resource shortages for mesoconsumers, thereby lim-iting their scope for antipredator behavior and indirectlyincreasing the rates at which they are killed by non-humanpredators. Nonlethal human activities, such as ecotourism,can cause animals to experience energetic and reproduc-tive costs that resemble the costs of predator avoidance.Notoriously, humans have eliminated top predators inmany systems, relaxing the need for mesoconsumers toinvoke antipredator behavior, potentially disrupting tro-phic cascades. As illustrated in the following sections,predation risk theory can provide tools for predicting,detecting, and potentially mitigating these and other con-servation problems.

366Encyclopedia of Animal Behavior

State-Dependent Risk-Taking: WhyHuman-Caused Resource Declines CanIncrease Predation Rates

During the 1980s, Marc Mangel, Colin Clark, JohnMcNamara, and Alasdair Houston pioneered models ofstate-dependent behavior (also known as dynamic statevariable models), which, among their many capabilities,predict the effect of residual reproductive value on behav-ioral decisions that maximize fitness. State-dependentrisk-taking, a subset of this theory, predicts that resourcedeclines and associated losses of body condition shouldincrease risk-taking and predation rates for individualsattempting to avoid imminent starvation or other netfitness losses. Bradley Anholt and Earl Werner providedearly empirical support for this prediction by experimen-tally exposing bullfrog tadpoles (Rana catesbeiana) to pre-dation risk from larval dragonflies (Tramea laceratea) undercontrasting levels of food abundance. Tadpoles experien-cing low food levels moved, on average, 1.5 times morefrequently and at higher speeds than tadpoles experien-cing high food levels. Higher movement rates underfood scarcity, which reflect greater foraging effort, alsoincreased exposure to predators and caused a 60% rise inmean predation rates, despite predator densities remainingconstant (Figure 1). This and later experiments by Anholtand Werner provided ground-breaking evidence that, con-sistent with the theory on state-dependent behavior, dichot-omous views about resource-driven (‘bottom-up’) versusconsumer-driven (‘top-down’) effects on population regula-tion are simplistic. Rather, these mesocosm experimentssuggested that synergisms between resources and predatorsare fundamental to population and community processes,and adaptive variation in prey behavior is inherent to thesesynergisms. Subsequent field studies suggest that thesesynergisms scale up to large vertebrates using vast land-scapes. Data on sea turtles, ungulates, and other large verte-brates suggest that individuals in poor body conditionspend more time in habitats with better food quality andhigher predation risk and consequently, may suffer higherpredation rates than individuals in better condition.

State-dependent risk-taking potentially has pro-found conservation implications because humans influ-ence the global distribution and abundance of resourcesused by animals. For example, models of state-dependentbehavior predict that overfishing can force harbor seals

(2010), vol. 1, pp. 366-376


2.0

Sw

imm

ing

spee

d (c

m s

–1)

(a)60

(b)

1.540

50

0.5

1.0

20

30

Per

cent

sw

imm

ing

0.0

10

0

60

High

High

HighLow

Low

Low

50

(c)

20

30

40

0

10

Per

cent

mor

talit

y

Food level

Figure 1 Early experimental evidence for how resource

declines may indirectly increase predation rates via behavioral

mechanisms of state-dependent risk-taking. (a) Meanpercentage (�1SE, N¼4 for each treatment) of observations

where bullfrog tadpoles were active in the presence of caged

predators (larval dragonfly) at high and low food levels. (b) The

mean speed of movement (�1SE, N¼ 4 for each treatment) ofactive bullfrog tadpoles in the presence of caged predators at

high and low food levels. (c) Mean percentage (�1SE, N¼20 for

each treatment) predation mortality of bullfrog tadpoles in the

presence of uncaged predators at high and low food levels. Datafrom Anholt BR and Werner EE (1995) Interactions between food

availability and predation mortality mediated by adaptive

behavior. Ecology 76: 2230–2234. Redrawn with permission fromBrad Anholt.

Conservation and Anti-Predator Behavior 367

(Phoca vitulina) and Steller sea lions (Eumetopias jubatus) inwestern Alaska to increase foraging effort at the cost ofincreased exposure to Pacific sleeper sharks (Somniosuspacificus) and other predators. Through these behavioralmechanisms, overfishing may indirectly increase predationrates, potentially contributing to declining trends for somepinniped populations (Figure 2).

Nonlethal Interactions Between Wildlifeand Humans

Humans interact with animals in many ways that arenonlethal yet potentially damaging. For instance, people

Encyclopedia of Animal Behavi

on foot or on motorized vehicles often approach wildlifeto the point of altering the animals’ activity and elicitingantipredator behavior like fleeing. These responses bywildlife to humans are referred to as ‘disturbance’ andtheir context includes ecotourism (e.g., wildlife viewing,photography), resource extraction (e.g., machinery use,blasting, helicopter access to remote sites), and the nonle-thal component of hunting (i.e., the search for quarry).Animal responses to disturbance stimuli may carry thesame cost of predator avoidance – lost opportunities forfeeding, mating, parental care, or other fitness-enhancingactivities – and often include the energetic costs of loco-motion while fleeing.

Theoretically, decisions by animals encountering humansshould follow the same principles as antipredator behav-ior: optimization of trade-offs between access to resources(or net energy gain) and avoidance of perceived danger.Consistent with this hypothesis, factors that influenceperceived risk of predation also affect animal responsesto disturbance stimuli. For instance, when approacheddirectly by a helicopter, Dall’s sheep (Ovis dalli dalli) far-ther from steep rocky slopes (a refuge from predation), fleesooner than sheep closer to these slopes (Figure 3).

Antipredator behavior and responses to disturbanceare analogous even when disturbance stimuli derive frommodern technologies that prey had not encountered previ-ously (e.g., motorized vehicles). This occurs because preyhave evolved antipredator responses to generalized threat-ening stimuli, such as rapidly approaching objects that crossa threshold of perceived safety. Early support for thishypothesis was provided by Larry Dill’s experiments inwhich zebra danios (Brachydanio rerio, a small fish) wereexposed to real predators, a predator-shaped model, and a‘cinematographic’ predator (a film of a black dot increasingin size, simulating an approaching object). In all cases, zebradanios fled when the angle subtended by the predator atthe prey’s eye reached a threshold rate of change (loomrate). The threshold loom rates depended on the size andspeed of the approaching ‘predator.’ Thus, danios appearedto decide when to flee by relating the loom rate to a marginof safety, regardless of whether the predator was real, amodel, or a film. Such generalized responses, however, arenot mutually exclusive with predator-specific responses.

Importantly, lack of overt response to disturbance sti-muli does not necessarily imply a lack of impact. Accord-ing to models of state-dependent behavior, an individual’sscope for antipredator behavior is influenced by the avail-ability of its resources, its current body condition, andother factors affecting residual reproductive value. Thus,animals lacking alternative sites with adequate resourcesor struggling to maintain adequate body condition may beunable to afford to abandon their resource patch and fleefrom disturbance stimuli. (The same principle applies tothe decision to abandon or care for dependent offspring.)People encountering wildlife often misinterpret lack of

or (2010), vol. 1, pp. 366-376


Shallow

Mid-depthDeep

(c)

(e)

(d)(h)(f)

(a)

(g)

(b)

0.6

0.8

1.0

4000Poor

Good

Seal body condition

0.2

0.0

0.4

2000

3000

Pro

port

ion

of fo

ragi

ngdi

ves

by s

eals

0

1000

Sharks

Pollock biomassHigh Low

Num

ber

of d

eep

dive

s by

sea

ls

0.8

1.0PollockHerring

30

40

50

0.2

0.4

0.6

Num

ber

of s

eals

kille

d by

pre

dato

rs

10

20

0.0

Sharks

Absent Present

Pro

port

ion

in s

eal d

iet

Pollock biomass

High Low0

Absent Present

Figure 2 Predictions from a model of state-dependent behavior on the relationship between fisheries and antipredator behavior

(arrows show causal links). (a) Pacific sleeper shark caught as bycatch in a trawl net targeting walleye pollock (Theragra chalcogramma)

in the Gulf of Alaska. On the one hand, the removal of top predators like sharks may relax predation risk and alter the behavior ofmesoconsumers, such as (b) harbor seals, thereby disrupting indirect effects of top predators on species at lower trophic levels. The

model predicts that shark removals will (c) greatly increase the proportion of dives to deep strata by seals, where both walleye pollock –

the seals’ most predictable resource – and sharks are most abundant. Consequently, shark removals should (d) greatly decrease andincrease, respectively, rates of seal-inflicted mortality on (e) Pacific herring (Clupea palassi) in shallow and mid-depth strata and

(f ) pollock in deep strata. (Note that Panels c and d represent only seals in good body conditions, which have greater scope for

antipredator behavior than seals in poor body condition.) On the other hand, fisheries depleting resources while top predators are still

present may increase state-dependent risk-taking and predation rates for mesoconsumers. When herring are not sufficiently abundantto compensate for pollock declines and sharks are present at a constant density, overfishing of pollock should increase for seals the

rates of (g) deep diving and, consequently, of (h) shark-inflicted mortality. Risk-taking and predation rates are exacerbated if seals are in

poor body condition (g, h). This modeling approach can explore net effects of (a) concurrent removals of resources and top predators.

Data are the outcome of computer experiments (N¼1000 forward iterations per treatment) simulating a 3-week winter period with themodel and protocols described in Frid A, Baker G G, and Dill L M (2008) Do shark declines create fear-released systems? Oikos 117:

191–201. Photo credits: (a) Elliott Lee Hazen; (b) Alejandro Frid, (e, f ) R.E. Thorne. In (b), the seal was captured for research purposes

and is about to be released carrying recording devices that measure diving behavior.

368 Conservation and Anti-Predator Behavior

fleeing as a neutral or even benign interaction. Yetbeneath this superficial appearance, animals may experi-ence decreased foraging efficiency due to increased vigi-lance, disrupted cycles of rest and digestion (which areparticularly important for ruminants), and physiologicalresponses to stress.

Similar to chronic risk effects from non-human pre-dators, it is theoretically plausible for chronic disturbanceto decrease long-term rates of energy intake and repro-ductive success. These risk effects can potentially leadto increased predation rates via mechanisms of state-dependent risk-taking, eventually causing populationdeclines (Figure 4). Similarly, it is theoretically plausible

Encyclopedia of Animal Behavior

for chronic disturbance stimuli to indirectly influenceplant community structure by causing herbivores tounderutilize plant resources in areas perceived to bedangerous while increasing use of plants in areas thatare perceived as safer.

Should animals not recognize that nonlethal stimulido not warrant the costs of antipredator behavior? Ani-mals rarely have perfect information and, theoretically,should maximize fitness by overestimating rather thanunderestimating risk. Overestimation costs, such as lostfeeding opportunities, are lower than underestimationcosts, which are death and loss of all future fitness. Thus,habituation to disturbance stimuli is only partial for

(2010), vol. 1, pp. 366-376


0.6

0.7

0.8

0.9

1.0

0 m 15 m

30 m

Distanceto rockyslopes =

0.2

0.3

0.4

0.5

Pro

babi

lity

of fl

eein

g

0.00.0(a)

(b)

0.5 1.0 1.5 2.0 2.5

0.0 0.5 1.0 1.5 2.0 2.5

0.1

Fle

dD

id n

otfle

e

Perpendicular distance from helicoptertrajectory (km)

Figure 3 (a) Estimated probabilities of Dall’s sheep fleeingduring helicopter overflights as a function of two independent

variables: the perpendicular distance (km) between the animal

and the helicopter’s trajectory (x-axis) and the animal’s distance

(m) to rocky slopes (a refuge from predation). Curves weregenerated from empirical data with logistic regression.

(b) Scatterplot corroborating estimates of fleeing probabilities;

dark circles represent sheep on rocky slopes, open circles

represent sheep 5–20m from rocky slopes (median¼ 20m), andcrosses represent sheep 25–1200m from rocky slopes

(median¼100m). Points are jittered so that overlapping data can

be read (i.e., there is no y-axis variability within response type).Note that the perpendicular distance from the helicopter’s

trajectory is a geometric correlate of angle of approach, with

smaller distances implying more direct approaches by the

helicopter. Both graphs suggest that sheep respond to themultiplicative (rather than additive) effects of the two

independent variables. Modified and reprinted with permission

from Frid A (2003) Dall’s sheep responses to overflights by

helicopter and fixed-wing aircraft. Biological Conservation 110:387–399.


nondomesticated animals. Even corvids and squirrels inurban parks – archetypal examples of habituation – maintainlevels of response to disturbance stimuli that are consistentwith principles of antipredator behavior.

Importantly, when habituation levels for prey surpassthose for predators, human infrastructure and its asso-ciated disturbance stimuli effectively provide prey withantipredator shields. For instance, Joel Berger found thatmoose (Alces alces) in the Yellowstone Ecosystem, USA,


generally avoid the vicinity of roads except during partu-rition. This choice of parturition sites may reduce calflosses to predation by grizzly bears (Ursus arctos), whichavoid roads more than moose do.

Relationships Between Risk Effectsand the Structure and Function ofEcosystems

Inceasing evidence suggests that top predators caninfluence the structure and function of ecosystems via acombination of direct predation and risk effects on theirprey. For example, the 1926 extirpation of wolves (Canislupus) from Yellowstone Natonal Park led to populationincreases and unrestrained browsing by elk (Cervus ela-phus) released from the lethal and risk effects of wolves.The combined population and behavioral changes by elklowered recruitment of woody riparian vegetation andupland deciduous trees. Since the 1995 reintroduction ofwolves, plant recruitment has improved partly becausedirect predation by wolves has reduced elk numbers andconsequently, the overall browsing pressure. Antipredatorresponses by elk, however, have influenced spatial varia-tion in the strength of the indirect effects of wolves onvegetation. This trophic cascade was strongest in riskierareas where elk foraged least: sites with abundant logs,particularly in riparian zones, where poor visibility andobstacles hindering escape may increase the probabilityof death by predation, given an encounter with wolves(Figure 5).

Trophic cascades mediated by antipredator behaviorcan modify physical habitat structures. For example, inZion National Park, unrestrained foraging by mule deer(Odocoileus hemionus) experiencing reduced predation riskfrom cougars (Puma concolor) appears to have erodedstream banks and reduced abundances of flora and faunaassociated with riparian zones. Similarly, in YellowstoneNational Park, the overgrazing of riparian trees andshrubs by elk released from wolf predation risk mayhave influenced the density and dam-building activitiesof beavers (Castor canadensis), which require woody plantsto construct dams. These correlational studies suggestthat, as mediated by ungulate herbivory and dammingby beavers, the loss of large carnivores has the potentialto indirectly alter hydrological processes that in turninfluence the structures and composition of aquatic andterrestrial communities. More generally, there is evidencethat herbivores across a wide range of body sizes andecosystems (e.g., from grasshoppers to marine turtles)can influence recruitment and community structure ofprimary producers, but risk and lethal effects of predatorslimit this influence.

It is plausible for risk effects also to influence thedynamics of nutrient flow across ecosystems. For example,

or (2010), vol. 1, pp. 366-376


Perceived risk

Antipredatorinvestment

(pre-energetic stress)

Net energy intake

Body condition

Antipredator investment(energetically stressed)

Reproductivesuccess

Predationrate

Population size

Increased rateof encounters withdisturbance stimuli

Increased rate ofpredator encounters

by prey

Figure 4 Conceptual model outlining the behavioral mechanisms by which increased rates of human-caused disturbance or

predator encounters by prey could cause population size to decline. Downward-facing arrows inside boxes indicate a negative

response and upward-facing arrows indicate a positive response. Modified and reprinted with permission from Frid A and Dill LM (2002)Human-caused disturbance stimuli as a form of predation risk. Conservation Ecology 6: http://www.consecol.org/Journal/vol6/iss1/

art11/print.pdf.


a recent study by Joseph Bump and colleagues indicatesthat moose on Isle Royale, Lake Superior, spend much oftheir time foraging on aquatic macrophytes and transferconsiderable amounts of aquatic nitrogen to terrestrialcommunities. Moose habitat use and, consequently, thespatial pattern and magnitude of their nitrogen depositionon land are influenced by wolf predation risk. The loss ofwolves could, therefore, alter the nutrient dynamicsof communities surrounding wetlands. Similarly, marinefishes that forage in seagrass ecosystems often shelter frompredators among mangrove prop roots or in reefs, therebytransporting seagrass-derived nutrients and energy intoreef and mangrove habitats. Declines of predators in seagrass system, therefore, can potentially alter the nutrientsubsidies into these habitats.

Mescocosm experiments by Oswald Schmitz provideevidence that, for invertebrate predator–prey interactionsin grasslands, the hunting mode of different predatorspecies (spiders) indirectly influences ecosystem function,as mediated by predator-specific antipredator responsesof herbivores (grasshoppers). Sit-and-wait ambush spidersleave persistent, point-source cues of predation risk. Con-sequently, grasshoppers can respond with chronic habitatshifts and their foraging pressure shifts from preferedplants associated with greater risk to less preferred plantsthat can be accessed more safely. As mediated by theserisk effects on grasshoppers, the indirect effect of sit-and-wait spiders on ecosystem function is greater plant speciesdiversity but lower primary productivity and nitrogenmineralization. In contrast, widely roaming spiders that


hunt actively do not leave predictable cues of predationrisk. Therefore, grasshoppers are unable to respond withchronic habitat shifts and roaming spiders have strongerlethal than risk effects on grasshoppers. As mediated bylower densities of grasshoppers, the indirect effect ofroaming spiders on ecosystem function is opposite tothat of sit-and-wait spiders: lower plant diversity butgreater productivity and nitrogen mineralization. Thesimilar nature of risk effects across diverse taxa and eco-systems suggests that these processes scale up and thatconserving species diversity within guilds of large preda-tors might be important to the maintenance of manyecosystem functions.

While the plight of top predators in terrestrial sysemshas long been recognized, only recently has it becomeapparent that marine ecosystems are experiencing cata-strophic losses of large predators through target andbycatch fisheries. These losses may create risk-releasedsystems mirroring those on land, where unrestrictedgrazing by herbivores or increased predation by mesocon-sumers affects the foundations of food webs. For example,studies in Shark Bay, Western Australia, reveal that taxaranging from herbivorous turtles and dugongs to piscivo-rous dolphins and seabirds modify their foraging locationsand behaviors to minimize risk from tiger sharks (Galeo-cerdo cuvier) and suggest that these antipredator behaviorsmay influence the structure of seagrass communities.Changes in seagrass communities in regions where tigersharks have been overfished further suggest that releasingmarine herbivores from predation risk may alter benthic

(2010), vol. 1, pp. 366-376

http://www.consecol.org/Journal/vol6/iss1/art11/print.pdf



Wolves

Elk

Browsing

Logs

With Without

Uplands

Riparian

Aspen height

1995 2000 2005

Year

0

50

100

150

200

250

0

20

40

60

80

100

0

20

40

60

80

100

120

0

5

10

15

20

Hei

ght (

cm)

Per

cent

bro

wse

dP

opul

atio

n (t

hous

ands

)P

opul

atio

n

(a)

(b)

(c)

(d)

Figure 5 Time series data suggesting trends that areconsistentwith the hypothesis that risk effects anddirect predation

from wolves indirectly affect aspen heights, as mediated by the

density and antipredator behavior of elk. (a) Wolf populations,(b) elk populations, (c) percentage of aspen leaders browsed, and

(d) mean aspen heights in Yellowstone’s northern range (early

springtime heights after winter browsing but before summer

growth). Direct predation by an expanding wolf populationdecreased elk numbers and contributed to decreasing browsing

pressure on aspen. Risk effects, however, influenced the spatial

variation in the strengthof this trophic cascade, as illustrated by the

stronger decrease in elk browsing and greater increase in aspengrowth in areas with downed logs, particularly in riparian zones,

where poor visibility and obstacles hindering escapemay increase

the probability of death by predation, given an encounter with

wolves. Caption adapted and figure reprinted (with permission)from Ripple W and Beschta R (2007) Restoring Yellowstone’s

aspen with wolves. Biological Conservation 138: 514–519.


communities. Losses of other large marine predators mayalso disrupt trophic cascades that are mediated by theantipredator behavior of mesopredators. Theoretical pre-dictions suggest that the removal of Pacific sleeper sharks


from northeastern Pacific ecosystems could shift pinnipedpredation from fishes in safer shallow waters to profit-able fish species in deeper waters that pinnipeds mightotherwise avoid to reduce predation risk (Figure 2). Justas in terrestrial systems, maintaining or restoring viablepopulations of large marine predators is likely importantto marine conservation.

Anthropogenic Climate Change andRisk Effects

Global climate change caused by human activities isaltering the resources, geographic distributions, phenol-ogy, and physiological and behavioral performance ofmany species at rates that exceed the range of naturalhistorical variability. Therefore, it can potentially createnovel ecological circumstances that attenuate or amplifyrisk effects or that indirectly increase predation ratesthrough state-dependent risk-taking (Figure 6).

Climate-Related Mechanisms ThatCan Attenuate Risk Effects

It is plausible that climate change indirectly attenuatesrisk effects in some systems by altering environmentalconditions that facilitate hunting success by predators.During winter on Isle Royale, the pack size and huntingsuccess of wolves on moose increases with snow depth.Historically, the North Atlantic Oscillation has drivensnow depth variation in the area. Under anthropogenicclimate change, however, the frequency of winters withshallow snow packs could increase and moose could poten-tially experience more frequent winters of relaxed preda-tion risk. Under this scenario, climate change wouldweaken the indirect effects of wolves on woody vegetation,as mediated by moose density and browsing behavior.

Climate change might also diminish the hunting effec-tiveness of some predators through physiological mechan-isms. Pursuit-diving seabirds and pinnipeds are endothermic(warm-blooded) and consequently, their burst speed whilehunting is unaffected by water temperature. In contrast, fish,their primary prey, are ectothermic (cold-blooded), andrising ocean temperatures are predicted to increase theirburst speeds and escape ability, potentially reducing therisk effects of seabirds and pinnipeds on fish.

Another mechanism potentially attenuating risk effectsis interspecific variation in climate-driven phenologicalresponses. Christiaan Both and colleagues have analyzedresponses to earlier spring warming by organisms at fourtrophic levels: deciduous oak trees (Quercus robur), cater-pillars of the winter moth (Operophtera brumata) that feedon oak buds, several species of passerine songbirds thatprey on caterpillars, and sparrowhawks (Accipiter nisus)

or (2010), vol. 1, pp. 366-376


Globalclimatechange

Resourceconsumption by

humans

Predator densityAvailable

resources/timingof reproduction formesoconsumers

Huntingeffectivenessof predators

Risk and lethal effectsof predators on

mesoconsumers

Community structureand ecosystem function

Figure 6 Simplified pathways illustrating hypothetical indirect effects of anthropogenic resource consumption and climate change on

community structure and ecosystem function, as mediated by mesoconsumer antipredator behavior. Solid and dotted lines represent,

respectively, direct and indirect links. Double pointed arrows indicate two-way relationships. To emphasize potentially dominant

mechanisms and maintain visual clarity, some plausible relationships are not shown (e.g., climate change may influence the localdensity of top predators and predation risk may alter the timing of reproduction). Whether human influences are predicted to attenuate

or amplify risk effects depends on ecological context (see text).


that prey on songbirds. Their findings indicate that overthe last two decades, oak budburst has been occurringprogressively earlier and moths are reproducing earlier,apparently tracking shifts in plant phenology. Similarly,songbird reproduction is occurring earlier, apparentlytracking moth reproduction. Sparrowhawks reproduc-tion, however, has not shifted to match the earlier peakabundance of songbird nestlings, which would enhancethe ability of sparrowhawk parents to provision theirnests. As underscored by this study, climate change mayindirectly decouple some predator–prey relationships,with mesoconsumers (e.g., songbirds) experiencing a sea-sonal lowering of risk effects from their predators (e.g.,sparrowhawks).

Climate-Related Mechanisms That Can AmplifyRisk Effects

Range shifts and expansions in response to climate changewill result in mesoconsumers encountering novel preda-tors, and animals that were top predators previously maybecome mesopredators. For example, warming ocean tem-peratures are predicted to facilitate the expansion to higherlatitudes of ectothermic sharks that are currently restrictedto lower latitudes. These range expansions could amplifyrisk effects for endothermic marine mammals and otherprey that currently live under lower predation risk fromsharks at higher latitudes.


Climate Change May Indirectly Alter MortalityRate Through State-Dependent Behavior

A general prediction derived from predation risk theory isthat climate change, which affects resource availability, canindirectly alter rates of mortality inflicted by predators (oranalogous agents: see below) through behavioral mechan-isms of state-dependent risk-taking. Data for diverse taxa –including amphibians, polar bears (Ursus maritimus), andhumans – are consistent with this prediction.

Efts (terrestrial juveniles) of the eastern red-spottednewt (Notophthalmus viridescens) compromise predatoravoidance when dry conditions force them to investmore on behaviors that relieve desiccation stress. Liveefts emit chemical cues signaling conspecific attraction,which facilitates huddling to reduce water losses. Recentlykilled efts, however, appear to emit a mixture of theseattractive chemical cues and of repulsive chemical cuessignaling predation risk. Experimental data indicate thatefts experiencing moist conditions avoid chemical cuesemitted by recently killed individuals, while these samecues attract efts stressed by desiccation. Drying trendsassociated with global climate change, therefore, canpotentially exacerbate predator-inflicted mortality ratesfor amphibians via these behavioral mechanisms.

Notably, climate-induced resource losses may forcetop predators to invoke foraging modes that increase therisk of human-caused mortality, creating a situationresembling that of reduced antipredator behavior by

(2010), vol. 1, pp. 366-376



energetically stressed mesoconsumers. Under normalconditions of sea ice, polar bears can meet energeticneeds by hunting for seals at breathing holes on the icepack and therefore, can avoid human settlements. In west-ern Hudson Bay, however, earlier break-up of sea iceduring spring has increased the period in which polarbears fast on land unable to access seals. Consequently,nutritionally stressed polar bears increasingly search forhuman-related food sources at settlements or camps,where humans kill them in self-protection. Sea ice breakup is predicted to keep shifting to progressively earlierdates, raising concern for polar bear conservation and themanagement of human–bear interactions.

Humans are not exempt from their own form of state-dependent risk-taking. Historically, climate-driven resourceshortages have influenced the decision by hungry societies toinitiate wars that might never have occurred had humanpopulations been well fed (Figure 7). Resource shortagesinduced by climate change can, consequently, exacerbate thepotential for war or other human conflicts. Arguably, this isthe biggest conservation issue of all, and – while not ignoringthe complexities of human cultures and modern social insti-tutions – principles of state-dependent risk-taking couldcontribute theoretical tools for anticipating and perhapsreducing conflicts between humans over scarce resources.

The converse of these examples is plausible whenclimate change is predicted to increase resource availabil-ity. For instance, Nicolas Lecomte and colleagues haveshown that snow geese (Chen caerulescens atlantica) in theArctic must leave their nests to drink. Drier conditionsincrease the dispersion of water sources and the durationof water acquisition trips, thereby amplifying nest lossesto predation. Climate models, however, predict increasedprecipitation for the Arctic over the next two decades,which can potentially enhance nest guarding and repro-ductive success by Arctic-nesting snow geese.

Managing Irrevocable Changesto the Biosphere

The conservation challenges ahead are daunting. Mostpressingly, climate change has already begun and willcontinue to reshape the biosphere. Encouragingly, a widerange of studies suggest that the basic mechanisms of riskeffects and state-dependent risk-taking have great general-ity. Thus, predation risk theory might have importantapplications to ecosystem conservation and the manage-ment of irrevocable changes to the biosphere. Some exam-ples are as follows.

Conservation Implications of State-DependentRisk-Taking

The conservation message of studies on state-dependentrisk-taking is that human-caused resource declines should


not be viewed solely as bottom-up impacts (e.g., nutri-tional stress), as is often the case. Instead, predictions andmitigation efforts should consider how the combinedeffects on resource availability of human consumption,nonlethal disturbance, and global climate change mightlimit the scope for antipredator behavior and potentiallyincrease rates of predation for many species. Conversely,human-caused resource subsidies may indirectly decreasepredation rates for some mesoconsumers, potentiallyaltering some predator–prey interactions.

The framework of state-dependent risk-taking can alsopredict the indirect influence of human-caused distur-bance stimuli on mortality rates inflicted by predators.Disturbance can functionally lower resource availabilitythrough increased vigilance or distributional shifts. Thus,wildlife managers might be able to predict scenarios inwhich chronic disturbance can increase energetic stress,thereby raising risk-taking while foraging and increasingpredation rates indirectly (Figure 4).

Managing Risk Effects of DisturbanceStimuli on Wildlife

Predation risk theory can provide a rationale for manag-ing disturbance stimuli without over-regulating humans.For instance, routes for motorized vehicles in remoteareas (e.g., helicopters used for resource extraction) maybe restricted to distances from known wildlife concentra-tions (e.g., raptor nests, ungulate birthing sites) that opti-mize the conflicting objectives of reducing disturbance towildlife while avoiding excessive detours. Similarly, pre-dation risk theory can be used to design wildlife viewingareas such that setback distances between people andanimals optimize viewing opportunities and preventionof disturbance, rather than merely stressing animals thatlack alternative habitats. Another concern is that huntingregulations generally consider only the lethal componentof hunting, yet hunters disturb many more animals thanthey kill. Thus, predation risk theory could help develophunting regulations that account for disturbance impactson targeted game. Further, when top predators areendangered but game species are not (e.g., Florida panther(Felis concolor coryi ) and white-tailed deer (Odocoileus virgi-nianus), respectively), regulations might also account forthe reduced hunting success that natural predators mightexperience because prey pursued by human huntersbecome more alert and difficult to capture.

Significantly, the use by prey of disturbance stimuli assafe zones that top predators avoid could diminish theeffectiveness of national parks for protecting biodiversity.Although national parks usually have the dual mandate offacilitating recreation and conservation, access for recrea-tion (e.g., roads, permanent campsites) might promote safezones for prey that disrupt predator–prey behavioralinteractions inherent to many aspects of biodiversity.

or (2010), vol. 1, pp. 366-376


1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900

Tem

p an

omal

y (�

C)

−0.15

−0.25

−0.35

−0.45

−5

45

30

15

0

33

22

11

0

20

15

10

5

0

1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 19000

9000000

18000000

27000000

36000000

45000000

25

50

75

100

125

15

35

55

75

Year AD

Pop

gro

wth

(%

)/fa

talit

y in

dex

No.

of w

ars

No.

of w

ars

Little ice age

(a)

(b)

(c)

(d)

Figure 7 State-dependent risk-taking by humans? Time series data on paleo-temperature variation, war frequency, and population

growth rate, AD 1400–1900, suggesting that climatic stress indirectly influences the willingness by humans to initiate conflicts and,

consequently, the rates of war-inflicted mortality (a prediction stemming from theory on state-dependent risk-taking). The mediating

factors between climatic stress and war (not shown here) are drops in agricultural production and rising food prices (see Figure 2 ofthe original source). (a) Temperature anomaly (�C) in the Northern Hemisphere that is smoothed by 40-year Butterworth low pass

filter. (b) Number of wars in the Northern Hemisphere (bright green), Asia (pink), Europe (turquoise), and the arid areas in the

Northern Hemisphere (orange). (c) Number of wars worldwide (colors represent estimates by different authors: see original source).

(d) Twenty-year population growth rate in Europe (turquoise), Asia (pink), and the Northern Hemisphere (blue) and the NorthernHemisphere 50-year fatality index (bright green). Cold phases are shaded as gray stripes. All war time series are in 10-year units.

The bright green curves correspond to the right y axis. Caption adapted and figure reprinted (with permission) from Zhang DD, Brecke P,

Lee HF, He YQ, and Zhang J (2007) Global climate change, war, and population decline in recent human history. Proceedings of the

National Academy of Science 104: 19214–19219.


Encyclopedia of Animal Behavior (2010), vol. 1, pp. 366-376



Predation risk theory could be applied to the design ofparks so that these safe zones for prey are predicted andmitigated.

Restoring and Conserving Risk Effects

A growing body of evidence suggests that the risk and directpredation effects of predators on mesoconsumers across awide range of body sizes and ecosystems can affect ecosys-tem function indirectly. If the influence of risk effects ontrophic cascades and related processes is as widespreadas these studies suggest, then conserving ecologicallymeaningful densities of top predators might be essential tomany aspects of community dynamics. As experiments byOswald Schmitz suggest, conserving predator diversity andits range of hunting modes, rather than merely conservingpredator abundance, may be required to maintain someecosystem functions (e.g., plant diversity, primary produc-tivity, and nitrogen mineralization in grasslands).

Where extinct already, restoring populations of upper-level native predators may potentially reverse manyaspects of ecosystem degradation via the reestablishmentof risk and lethal effects; wolves reintroduced to Yellow-stone provide a compelling case. In terrestrial systemswhere predator reintroductions are impossible or wherehuman-dominated landscapes no longer support the habitatrequirements of large carnivores, hunting can potentiallybecome a management tool for reducing some ecologicalconsequences of the loss of direct predation and risk effectson mesoconsumer populations. This potential, however, hasyet to be met because hunting regulations and access logis-tics (e.g., the distribution of roads or navigable waterways)typically limit human hunting behavior to spatiotemporaldistributions and patterns of prey selectivity that differ fromthose of non-human carnivores. Predation risk theory, com-binedwith demographic analyses, could be used to optimizehunting regulations so that human hunters mimic some riskand lethal effects of non-human carnivores more closelywithout threatening human safety and prey populationsthemselves. Carnivore restoration and hunting as manage-ment tools, however, are difficult to reconcile with somevalues of human society, which could preclude some appli-cations of predation risk theory.

A related concern is that predator reintroductionscould cause high predation rates on prey that may havelost their antipredator skills after living under relaxedpredation pressure for several generations. Data forsome vertebrate taxa (e.g., moose, passerine birds, marsu-pials), however, suggest that predator-naı̈ve prey can learnto recognize and avoid novel predators within one gener-ation. Reintroduced predators, therefore, may have theirhighest predation rates on naı̈ve prey early in their geo-graphic expansion, but their kill rates may diminishas prey become savvier. Although the ability to learnantipredator skills does not negate concern that novel


predators could drive extinct small populations of naı̈veprey, this issue generally applies to exotic predators invad-ing small islands rather than to planned reintroductions ofnative carnivores. A related problem is that efforts torestore native mesoconsumers (e.g., endangered marsu-pials in Australia) through translocations may fail unlesscaptive-raised mesoconsumers are trained to recognizeand avoid predators before being released.

Marine top predators and their risk effects have beenlargely overlooked until recently. There is growing recog-nition that marine top predators like sharks, classic vil-lains in popular culture, may have important indirecteffects on marine communities, as mediated by the anti-predator behavior of their prey. Sharks and other marinetop predators have been declining almost worldwidebecause of target and bycatch fisheries. Their conserva-tion and restoration may be essential to many aspects ofmarine ecosystem function, and predation risk theory canbe used to support arguments for modifying fishery quo-tas and establishing marine reserves accordingly.

Although major losses of ecological integrity related tohuman resource consumption, climate change, andrelated processes will be inevitable, there is cautiousoptimism that predation risk theory can help predict andpotentially reduce some of the damage. The maintenanceof antipredator behavior over large ecological scales couldwell be a litmus test for our ability to conserve manylevels of biodiversity.

Acknowledgments

We thank the following people for reviewing earlierdrafts and helping us improve the manuscript: LarryDill, Diana Raper, Anne Salomon, Brooke Sargent, andTed Stankowich.

See also: Anthropogenic Noise: Implications for Conser-

vation; Conservation and Behavior: Introduction; Ecology

of Fear.

Further Reading

Berger J (2007) Fear, human shields and the redistribution of prey andpredators in protected areas. Biology Letters 3: 620–623.

Both C, van Asch M, Bijlsma R, van den Burg A, and Visser M (2009)Climate change and unequal phenological changes across fourtrophic levels: Constraints or adaptations? Journal of Animal Ecology78: 73–83.

Cairns D, Gaston A, and Huettman F (2008) Endothermy, ectothermyand the global structure of marine vertebrate communities. MarineEcology Progress Series 356: 239–250.

Caro TM (ed.) (1998) Behavioral Ecology and Conservation Biology.New York, NY: Oxford University Press.

Clark CW and Mangel M (2000) Dynamic State Variable Models inEcology. New York, NY: Oxford University Press.

Creel S and Christianson D (2008) Relationships betweendirect predationand risk effects. Trends in Ecology & Evolution 23: 194–201.

or (2010), vol. 1, pp. 366-376



Festa-Bianchet M and Apollonio M (2003) Animal Behavior and WildlifeConservation. Washington, DC: Island Press.

Frid A, Baker GG, and Dill LM (2008) Do shark declines createfear-released systems? Oikos 117: 191–201.

Frid A and Dill LM (2002) Human-caused disturbance stimuli as a form ofpredation risk. Conservation Ecology 6. http://www.consecol.org/Journal/vol6/iss1/art11/print.pdf.

Heithaus MR, Frid A, Wirsing AJ, and Worm B (2008) Predictingecological consequences of marine top predator declines. Trends inEcology & Evolution 23: 202–210.

Knight RL and Gutzwiller KJ (eds.) (1995) Wildlife and Recreationists:Coexistence Through Management and Research.Washington, DC:Island Press.

Lecomte N, Gauthier G, and Giroux J (2009) A link between wateravailability and nesting success mediated by predator–preyinteractions in the Arctic. Ecology 90: 465–475.

Ray JC, Redford KH, Steneck RS, and Berger J (eds.) (2005)Large Carnivores and the Conservation of Biodiversity.Washington, DC: Island Press.


Ripple W and Beschta R (2004) Wolves and the ecology offear: Can predation risk structure ecosystems? Bioscience54: 755–766.

Rohr JR and Madison DM (2003) Dryness increases predation risk inefts: Support for an amphibian decline hypothesis. Oecologia 135:657–664.

Schmitz OJ (2008) Effects of predator hunting mode on grasslandecosystem function. Science 319: 952–954.

Schmitz OJ, Grabowski JH, Peckarsky BA, Preisser EL, Trussell GC,and Vonesh JR (2008) From individuals to ecosystem function:Toward an integration of evolutionary and ecosystem ecology.Ecology 89: 2436–2445.

Schmitz OJ, Post E, Burns C, and Johnston K (2003) Ecosystemresponses to global climate change: Moving beyond colormapping. Bioscience 53: 1199–1205.

Stirling I and Parkinson CL (2006) Possible effects of climate warming onselected populations of polar bears (Ursus maritimus) in theCanadian Arctic. Arctic 59: 261–275.

(2010), vol. 1, pp. 366-376



this article was originally published in the encyclopedia ......encyclopedia of animal behavior...

Documents