US Department of Agriculture US Government Publication Animal and Plant Health Inspection Service Wildlife Services

Costs and effectiveness of damage management of anoverabundant species (Sus scrofa) using aerial gunning

Amy J Davis AC Bruce LelandB Michael BodenchukB Kurt C VerCauterenA

and Kim M PepinA

ANational Wildlife Research Center Wildlife Services Animal Plant Health Inspection ServiceUnited States Department of Agriculture 4101 Laporte Avenue Fort Collins CO 80521 USA

BWildlife Services Animal Plant Health Inspection Service United States Department of AgricultureSan Antonio TX 78269 USA

CCorresponding author Email amyjdavisaphisusdagov

AbstractContextManagement of overabundant or invasive species is a constant challenge because resources for management

are always limited and relationships between management costs population density and damage costs are complex anddifficult to predict Metrics of management success are often based on simple measures such as counts which may not beindicative of impacts on damage reduction or cost-effectiveness under different management plans

Aims The aims of this study were to evaluate the effectiveness of aerial gunning for the management of wild pigs(Sus scrofa) and to evaluate how cost-effectivenesswould vary under different relationships between levels of damage anddensities of wild pigs

Methods Repeated reduction events were conducted by aerial gunning on three consecutive days at three study sitesUsing a removal model the proportion of the population removed by each flight was estimated and population modellingwas used to show the time it would take for a population to recover Three possible damagendashdensity relationships werethen used to show the level of damage reduction (metric of success) from different management intensities and levelsof population recovery and these relationships were expressed in terms of total costs (including both damage andmanagement costs)

Key results Populations were typically reduced by ~31 for the first flight ~56 after two flights and ~67 after threeflights When the damage relationship suggests high damage even at low densities the impact of one two or three flightswould represent a reduction in damage of 2 19 and 60 respectively after 1 year Different damage relationshipsmay show considerable damage reduction after only one flight Removal rates varied by habitat (005 per hour in openhabitats compared with 003 in shrubby habitats) and gunning team (003 versus 005)

ConclusionsMonitoring the efficacy of management provides critical guidance and justification for control activitiesThe efficacy of different management strategies is dependent on the damagendashdensity relationship and needs further studyfor effective evaluation of damage reduction efforts

Implications It is critically important to concurrently monitor density and damage impacts to justify resource needsand facilitate planning to achieve a desired damage reduction goal

Additional keywords costndashbenefit analyses damage reduction feral swine invasive species monitoring populationreduction wildlife management

Received 21 November 2017 accepted 5 October 2018 published online 18 December 2018

Introduction

An important element in any wildlife management programincluding for invasive species is monitoring the state of thesystem (Hauser et al 2006Lyons et al 2008 Fackler andHaight2014) to allow for evaluation and refinement of managementactions To establish meaningful metrics of managementsuccess it is important to define the management objectiveclearly For many game species the objective is to maintain aviable population while keeping the population under a

threshold that minimises humanndashwildlife conflicts anddamage (Diefenbach et al 1997 Fagerstone and Clay 1997VerCauteren et al 2011) In these situations changes inabundance provide an evaluation of the efficacy ofmanagement (Diefenbach et al 1997 Rupp et al 2000) Incontrast for invasive species or overabundant populationswhere the objective is to reduce damage (Conover 2001Campbell and Long 2009 Bleier et al 2012) changes indamage levels are the most important metric of management

CSIRO PUBLISHING

Wildlife Research 2018 45 696ndash705httpsdoiorg101071WR17170

Journal compilation CSIRO 2018 wwwpublishcsiroaujournalswr

success (Hone 1996) When the underlying objective is cost-effectiveness then the impact on the population the damagecosts and costs for removal actions need to be combined to makethe best decision

A common method for monitoring managed wildlifepopulations is using count data which can be harvest counts(Rupp et al 2000 Linnell et al 2010) management take counts(Slade and Blair 2000) or index survey counts (DeMaso et al1992 Bengsen et al 2011) Many advancements have beenmade in how these types of count data are analysed including(but not limited to) population reconstruction for harvest data(Gove et al 2002) removal models for take data (Zippin 1958)and distance sampling for count surveys (Buckland et al2001) When populations are being actively controlled takedata from management actions are routinely collectedRemoval models provide estimates of initial abundanceproportion removed remaining abundance and removal rate(Farnsworth et al 2002 St Clair et al 2012 Davis et al2016) However removal models require multiple removalevents within a period of demographic closure which is oftennot how management activities are conducted In these casesthe metrics used to measure short-term success of managementmay not be indicative of long-term success (Shrnhur et al1997) making it challenging to determine the probability oflong-term success When counts of take are the measure ofsuccess for short-term management managers are incentivisedto focus efforts in high-density areas and shift resources whenbody counts decline because it requires less effort to removelarge numbers of individuals when population density is highTherefore conducting multiple removal events may not bethought of as cost-effective when only management costs areconsidered because management costs increase as populationdensities decline (Baxter et al 2008) Thus counts of individualsremoved could provide some measure of management successwhen population densities are high but provide less informationwhenpopulationdensities are lowand on their own donot allowfor assessment of whether a particular damage reductionobjective was met

Population growth rates and models can be used to projectpopulation size in time and thus evaluate the time it takes for apopulation to rebound after a reduction event ndash the long-termimpacts of management If population management is sporadicor continuous the rebound timing gives a sense of the level ofmanagement needed to make a lasting impact on the populationAn understanding of long-term population dynamics allowsfor prediction of the effects of different reduction efforts onlong-term damage reduction In addition to potential non-linearimpacts of population reduction on long-term managementsuccess the impact of different population densities mayrelate non-linearly to wild pig damage (Hone 2002) Damageis a complex and variable metric to estimate but is importantto quantify for ensuring that economic resources are beingallocated efficiently

Wild pigs (Sus scrofa) are a global pest species and aninvasive species in the Americas and Australia (Barrios-Garcia and Ballari 2012 Bengsen et al 2014 Bevins et al2014) Wild pig populations have grown and their range hasexpanded over the last three decades in the United States ofAmerica (Bevins et al 2014 Snow et al 2017) Increasingly

resources have been applied to slow the spread of wild pigs andreduce the damage to crops infrastructure and other species(Seward et al 2004 Campbell and Long 2009 Bevins et al2014) Many methods are currently in use to reduce wildpig population sizes including aerial gunning (Campbell et al2010 Parkes et al 2010) trapping and snaring (Choquenot et al1993 Caley 1994 Williams et al 2011) hunting (Naughton-Treves 1998 Waithman et al 1999) and new methods suchas toxicants and fertility treatments are being developed(Cowled et al 2008 Massei et al 2008 Snow et al 2017)Aerial gunning is known to be an efficient method for removingwild pigs (Choquenot et al 1999 Parkes et al 2010 Davis et al2017) While it is well documented that costs of removing pigsby aerial gunning increase dramatically as population decreases(Choquenot et al 1999) its cost-effectiveness in terms ofdamage reduction is unknown

The relationship between damage and density is not wellunderstood particularly for wild pigs Wild pigs cause anenormous amount of damage in the United States of Americawith one estimate around USD15 billion in damages per year(Pimentel 2007) There is considerable effort being employedto reduce the damage caused by this invasive speciesUnderstanding the relationship between wild pig densities andthe amount of damage they inflict is critical to reducing damageefficiently andcost-effectivelybecause it helps informhowmanypigs need to be removed to reach a desired damage level

To improve the ability to predict cost-effectiveness of wildpigmanagementprogramsweanalyseddata fromaerial gunningof wild pigs at three sites in Texas USA We examined effectsof three damagendashdensity relationships on potential costs andeffectiveness of the wild pig damage management work to giveexamples of how these factors will influence the decision matrixof managers Our specific objectives were to (1) estimate theeffectiveness and costs of repeating aerial gunning over thesame space ndash both in the short and long-term (2) quantifyfactors that affect effectiveness of aerial gunning and (3)examine the effects of different wild pig densityndashdamagerelationships on effectiveness and costs

Methods

We conducted wild pig reductions in three habitats in TexasSite 1 was an open semiarid desert habitat along the Pecos Riverin Culberson Reeves Loving and Ward counties Most of the850-km2 site consisted of shrub land and barren land withless than 2 of either woody wetlands or cultivated cropsimmediately around the Pecos River (Homer et al 2015) Site1 was in the trans-Pecos portion of the state and removal workwas conducted in this area toprevent the spreadofwild pigs northandwest andprevent future damage rather than to reduce currentdamageThus damage reductionwasnot adirect objective at thissite Site 2was a 105-km2 area in dense cedar and live-oak habitatin Burnet County Approximately 65 of Site 2 was forested~32was shrub and grassland with lt3 developed open spaceand openwater (Homer et al 2015) Site 3was a 122-km2 area ina coastal plain habitat inMatagordaCounty This site hadwoodywetlands as a dominant feature (49) 20 was pastureland15was shrub and grassland and ~7 dense hardwood (Homeret al 2015) In Sites 2 and 3 damage reduction was the main

Cost effectiveness of damage management Wildlife Research 697

objective All study areas are on private land andwere conductedat the request of the landowner

We used aerial gunning from helicopters to remove wild pigsfrom each study area Aerial gunning was conducted by theUnited States Department of Agriculture (USDA) Animal andPlant Health Inspection Service (APHIS) Wildlife Services(WS) which provides wildlife control assistance to landowners based on the authority of the Animal Damage ControlAct of 1931 (7 USCA x 8351 ndash 8354) and in compliance with theNational Environmental Policy Act (42 USC x 4321) Theobjective was to remove (kill) all pigs encountered the searchstrategy used was based on the expert opinion of the pilotand gunner team and included searching the entire study areaFlights over Site 1 were conducted in April 2016 flights overSite 2 were conducted in June 2016 and flights over Site 3 wereconducted in July 2016 Three flights per site were conducted onsubsequent days We had two pilot and gunner teams (A and B)Team A had 10ndash12 years of experience conducting aerialgunning and Team B had 5ndash7 years of experience conductingaerial gunning In Site 1 TeamAflew on the first day then TeamBon the secondday and thenTeamAagainon the third dayThiswas the same for Site 3 but the order was switched for Site 2Flights were conducted with minimal winds (considerably lessthan the visual flight rules requirements of 26 kn) and noprecipitation Total flight times were generally between 6 and8 h per day excluding time on the ground for refueling Twelve-gauge shotgunswere usedwith double-aught buckshotCarewastaken to ensure all shot animals were killed (animal was seen tocollapse and was often double tapped ndash two shots fired in rapidsuccession at the same target and shot animals were flown backover to verify theywere down beforemoving on) The number ofanimals removed and the hours in the helicopter were recordedfor each flight

We estimated abundance from the aerial gunning datausing a removal model that simultaneously estimatesabundance and removal rates from counts of removedindividuals (Farnsworth et al 2002 Royle and Dorazio 2006)We implemented the removal model in a hierarchical Bayesianframework that accounted for variation in removal effort (Daviset al 2016) Removal models require multiple reduction eventswithin a time period in which there is demographic closure(no births deaths immigration and emigration) Removalestimators are advantageous for monitoring abundance ofinvasive species because they directly use data from controlevents (Zippin 1958 Farnsworth et al 2002 Chee and Wintle2010) for estimation of abundance pre-control From theestimates of abundance pre-control and numbers ofindividuals removed the proportion by which the populationis reduced due to control can be derived We also incorporatedcovariates on removal rate specifically personnel and habitat(see SupplementaryMaterial 1 for model details) We comparedthe removal rates for the two pilot and gunner teams andexamined the relationship with the proportion of forest cover(as determined from Homer et al 2015) We estimated thepopulation remaining after each flight by subtracting thenumber removed by flight from the estimate of initialabundance We then derived the proportion of the populationremoved by each flight by dividing the number removed by eachflight by the estimated population size pre-flight

To examine the long-term impact that reductions had onpopulations we used a standard logistic population growthmodel (Sibly and Hone 2002 Otto and Day 2007) to predictabundance at 1 3 and 5 years following a reduction event Weused estimates of population growth rate and density dependencefrom previous studies to provide the boundaries for ourmodelling (Caley 1993 Bieber and Ruf 2005 Timmons et al2012 Mellish et al 2014) Previous studies on wild pigpopulations have estimated growth rates ranging from 085 to163 (Bieber and Ruf 2005) with the majority of growth-rateestimates between 115 and 135with amean of 124 (Bieber andRuf 2005 Timmons et al 2012 Mellish et al 2014 Moretti2014) We examined a range of growth rates (l) from 1 to 2 toprovide a broader picture of population growth rates that may beexhibited

Management programs for wild pigs are often aimedat reducing damage caused by wild pigs Damage costs canrange from quantifiable assessments of lost crops infrastructurerepairs and livestock feed loss to more abstract costsassociated with disease risks landscape rooting and predationor competition of native imperilled or game species Damagecosts are a combination of all of these factors however there areareas with little or no current damage costs but with the potentialfor great damage if wild pigs are able to establish a population inan area Therefore damage reduction can be in current costs aswell as the prevention of future costs

The relationship between wild pig population densities anddamage incurred bywild pigs is not well known but studies havesuggested that the damagendashdensity relationship for species ingeneral is not necessarily linear (Hone 1995West and Parkhurst2002) To demonstrate the importance of knowing the shape ofthe damagendashdensity relationship for wild pigs we show theimpacts of damage under three potential relationships (Fig 1)linear (damage and density are linearly related) saturating (even

100

Saturating Linear Exponential

100

Population

D

amag

e

80

80

60

60

40

40

20

20

0

0

Fig 1 Theoretical relationships between the percentage of a populationremaining and the percentage of total damage remaining

698 Wildlife Research A J Davis et al

at low levels of density the damage is high) and exponential(only at high levels of density is the damage high)We comparedpercentage damage reduction after 1 3 and 5 years by varyinglevels of percentage population reduction (ranging from 0to 100 reduction) and growth rate (l= 1ndash2) under the threedamage relationships

The damage was not explicitly estimated directly at the sitesin our study Therefore we examined hypothetical damage coststo examine the costndashbenefits of different levels of managementeffort Damage can be to crops pastureland or infrastructureIn our study areas the damage was on pastureland The costsmay depend on the type of damage so we examined a range ofvalues based on average values for different land types peracre and different numbers of acres The range is flexible($0ndash$150 000) to account for different types of damage anddifferent sizes of properties We used the range of damage coststo compare the relative damage costs and management costsof wild pigs across a range of percentage population reductions(0 to 100) Estimated costs of helicopter removal work wasbased on the hourly rate charged for the reduction work weconducted (USD625 per hour)

Results

We removed 324 147 and 362 wild pigs from Sites 1 2 and 3respectively across all flights conducted The total numberremoved by flight and the time spent in the helicopter areshown by site and flight in Table 1 We estimated initialabundance at Sites 1ndash3 to be 413 (95 credible intervals ndash

CI 297 ndash 481) 268 (95 CI 200ndash318) and 545 (95 CI387ndash646) respectively (Fig 2) The average proportion of thepopulation removed after the first flight was 31 (se = 6)cumulatively after the second flight was 56 (se = 15) andcumulatively after the third flight was 67 (se = 12)

Removal rates differed based on the pilot and gunner team(Fig 3A) In our example Team A had a lower hourly removalrate (003 95 CI 002 004) than Team B (005 95 CI 004006 Fig 3A) We found removal rate decreased as percentagecover increased (b= ndash055 95 CI ndash079 ndash027 Fig 3B)

Growth rates impacted the linear damage relationshipmore substantially than either the saturating or the exponentialrelationships (Fig 4) At slower growth rates there was still

Table 1 Raw data from pilot study removing wild pigs from threestudy sites in Texas from aerial gunning work by site

Habitat Pilot andgunner team

Flight No wildpigs removed

Hours

Site 1 Team A 1 157 8Team B 2 140 8Team A 3 27 67

Site 2 Team B 1 71 75Team A 2 43 75Team B 3 33 8

Site 3 Team A 1 161 7Team B 2 125 69Team A 3 76 59

700

600

500

400

300

200

100

38

72

78

2643

55

30

52

66

0Site 1 Site 2 Site 3

Pass 3Pass 2Pass 1Initial

Abu

ndan

ce

Fig 2 Abundance estimate (with 95 confidence intervals for the initialabundance) by site for the initial population size and the subsequent size aftereach flight The proportion of the population removed is shown in text byflight above the abundance estimate

004

003

005

(a) (b)

Rem

ovea

l rat

e

002

Team A Team B 0 20

Site 1

Site 3

Site 2

40 60 80 100

Cover

004

002

000

Fig 3 (A) Boxplot of the posterior distributions of hourly removal rates by pilot and gunner team The greyboxes represent the interquartile range the dotted lines show the 95 credible intervals and the circles representthe total rangeof the posterior distributions (B)Relationship of hourly removal rate and thepercentage cover theshaded region represent the 95 credible intervals Points shown represent the percentage forest cover by site

Cost effectiveness of damage management Wildlife Research 699

damage reduction for all densityndashdamage relationships after5 yearswhen 40or less of the populationwas removed (Fig 4)

The hours of aerial gunning required grew exponentially asthe proportion of the population removed increased (Fig 5A)When the hourly removal rate was greater than 01 the effortrequired to remove up to 90 of a population was reasonable(lt20 h offlight time Fig 5B) If the hourly removal ratewas 004(as was the average in our study) the effort needed to removemore than 90 of the population was more than 45 h (Fig 5A)

The relationship between the reduction in damage costsand removal costs is an important factor in evaluating thecost-effectiveness of management actions For a saturatingdamagendashdensity relationship the population needed to bereduced substantially (80 or more) for the damage reductioncosts to outweigh the removal costs (Fig 6A) The total costsfor the linear relationship and the exponential relationshiponly showed high overall costs when the initial damage costswere high and the population reduction was low (Fig 6B C)

We compared the overall costs with the initial damage costs todetermine if the resulting removal effort resulted in an economicgain or loss over doing nothing For the saturating relationshipwe found that damage costs must be greater than USD50 000and more than 60 of the population must be removed to see aneconomic gain over baseline damage costs (Fig 6D) If thedamage relationships were either linear or exponential therewas generally a cost saving to removing at least some portion ofthe population for almost all levels of damage (Fig 6E F)

Discussion

For other ungulate species the damagendashdensity relationship hasbeen estimated to be high even at low densities ndash similar to thesaturating relationship we presented (Gill 1992 Hone 2002West and Parkhurst 2002) Under the saturating relationshipwe found that the proportion of the population that needs to beremoved in order to have a substantial effect on damage can be

20

18

16

14

100

80

60

40

20

0

12

10

20

18

16

Gro

wth

rat

e

Afte

r 3

Year

sA

fter

5 Ye

ars

Afte

r 1

Year

14

12

10

20

18

16

14

12

1020 40 60 80 20 40 60

Population

Saturating Linear

Damage-density relationshipsExponential

80 20 40 60 80

100

o

f Dam

age80

60

40

20

0

100

80

60

40

20

0

Fig 4 Percentage of total damage (as calculated without any wild pig reduction) by the percentage of the initialwild pig population (eg 20 population corresponds to an 80 reduction in wild pigs) and potential values ofwild pig growth rates The percentage of total damage is shown cumulatively for three time frames (1 3 and 5 yearspost reduction event shown in rows) and for three different potential damagendashdensity relationship (saturating linearand exponential shown in columns)

700 Wildlife Research A J Davis et al

greater than 80 If a saturating relationship is representativeof the damagendashdensity relationship for wild pigs our resultssuggest that it may be more cost-effective to conduct repeatedremoval events within a short time frame than to spread thesame effort over a longer period or a larger area especially at

high-damage sites In contrast if the densityndashdamagerelationship is linear or exponential substantial damagereduction can occur by reducing the population by 50 oronly 20 respectively both of which are substantiallydifferent levels of management effort Thus empirical work

100

(a) (b)

100

005

Population reduced

010

Rem

oval

rat

e

015

020

80

20

20

0

0

60

6040 100

Effo

rt (

eg

hou

rs in

a h

elic

opte

r)

25 +

20

15

10

5

80200 6040

Effo

rt r

equi

red

(in h

ours

)

Fig 5 (A) The effort required is plotted against the percentage of the population removed based on a removalrate for one hour of effort of 004whichwas estimated fromour data (B) Reduction effort (hours in a helicopter)necessary to remove varying proportions of the population (x-axis) by the removal rate given one hour of effort(y-axis)

150 000

(a) (b) (c)

(d ) (e) (f )

100 000

50 000

Tota

l cos

tsC

ost s

avin

gs

0

150 000

100 000

50 000

0

150 000

ndash150 000

100 000

ndash100 000

50 000

ndash50 000

0

150 000

100 000

50 000

000 02 04 06 08 10 00 02 04 06

Proportion population removed08 10 00 02 04 06 08 10

Saturated Linear Exponential

Fig 6 Estimated total costs (damage costs plusmanagement costs top row) andcost savingsover doingnothing (bottom row) for a rangeofinitial damage costs (y-axes) and proportions of the population removed (x-axes) under three damagendashdensity relationship scenariossaturated (AandD) linear (BandE) andexponential (CandF) The total costs are shownranging from lowcosts (blue) tohighcosts (red)Thecost savings are shown inwhitewhen the costs are similar to the initial damage costs blue for cost savings and red for additional costs over theinitial damage costs

Cost effectiveness of damage management Wildlife Research 701

to determine the shape of this relationship is needed formanagersto estimate the most cost-effective actions for a particular levelof damage Once this relationship is better understood we cancompare actual management costs to damage costs

The damagendashdensity relationship may not be constant acrossspace and time and management efforts are often conductedunder the assumption that the damage reduction is greater atdifferent times of year For instance the relationship betweendamage and density may be saturating or linear at some times ofthe year but exponential at others Managers may time removalevents to occur immediately before planting or harvesting ofspecific crops such as peanuts because these crops are onlytargeted by pigs at those times of the year In these cases doingonly two removal events ndash one right before planting and oneright before harvest ndash may be the most cost-effective damagereduction strategy compared with conducting several flightsin the middle of the growing season Additionally theremoval rate is known to vary at different times of the yeardue to weather and seasonal deciduous canopy cover Thereforein addition to the fluctuating potential to reduce damagethe cost-effectiveness may change throughout the year due tochanges in the costs of removing individuals These are standardconsiderations that managers account for when planningremoval work and are necessary to account for in models formanagement effectiveness

We have provided a simplified version of how adamagendashdensity relationship may look and shown howimportant that relationship can be in connecting managementto the objective of damage reduction Single removal events canbe beneficial for many reasons even if the cost calculations donot suggest the reduction in damage is greater than the cost ofremovals Aerial efforts may be employed to access sites thatare not feasible to access from ground efforts (eg trappingor ground shooting) or to target areas that are particularlyvulnerable to damage (eg predation of livestock growthstage of crops) Additionally the damagendashdensity relationshipmay be more complex because different age classes or thedifferent sexes might have a different impact on the damageincurred While young pigs may be more influential onpopulation growth a single adult boar can cause a substantialamount of damage to the landscape Thus it may be mosteffective to use a removal strategy that targets the subset ofthe population that has the largest effect on the underlyingobjective (eg long-term population reduction or immediatedamage reduction)

Reducing wild pig populations has the potential benefit ofreducing damage costs however the reduction events are notwithout costs themselves As a larger portion of the populationis targeted for reduction the costs of reduction increaseexponentially (our results and Choquenot et al 1999) Thecosts to implement reductions in populations can be comparedwith the potential cost savings of damage reductions whichmay help justify expenses for removal work Examining thesecost trade-offs up front can be considerably beneficial forbudget planning In particular population projections can helpjustify the need for large up-front contributions to managementefforts because these large initial investments may result inconsiderably lower long-term costs In contrast consistentsmaller investments in removal work over time may have the

samemanagement costs as one initial intensive reduction effortbut the long-term damage costs may be considerably higheras these may be simply lsquomowing the grassrsquo and not pushing apopulation towards elimination

The damagendashdensity relationship heavily influences howmuch population reduction is necessary to induce a costsavings Generally if the relationship is linear or exponentialdoing almost any removal results in a cost savings unless thereare little or no damage costs Additionally it is important toconsider that reducing future damage costs by reducingthe spread of the species to new areas and preventing damagefromoccurring may result in future savings Site 1 is an examplewhere the primary objective was to prevent the spread of wildpigs on the border of their range in Texas to prevent futuredamage in neighbouring areas Thus even if current damagecosts are low anticipated future damage costs may warrantan immediate expenditure Our results on the relationshipbetween damage costs and population density could be usedeither for planningmanagement against current damage levels oranticipated damage levels However the cost savings we showare short-term savings not accounting for potential effects ofmanagement on suppressing geographic spread which couldlead to additional cost savings that we have not incorporatedpresenting a challenge for future work There is considerablevalue in examining these potential future costs that may resultif wild pigs are allowed to spread to new areas An actualaccounting of the theoretical damage to a new area may makea strong case to focus removal efforts in areas with little currentdamage but that are on the edge of an invasion front instead ofin areas with high current damage since the long-term damagemaybe considerably higher if the invasion is allowed to continue

Repeated reduction events allowed us to estimate theabundance of wild pigs in each of our study areas whilesimultaneously reducing their numbers In our study areas wereduced the populations by an average of 31 after one flightand an average of 67 by the completion of three flights Ifwe had not conducted multiple removals and thus were not ableto estimate abundance we would only have known that we hadremoved 157 71 and 161 individuals in the respective sites(after one flight) In this case we may have expected that wehad impacted Sites 1 and 3 similarly because they had a similarnumber of animals removed when in fact we had removedalmost 40 of the population in Site 1 and only 30 of thepopulation in Site 3 If we were to assume that both populationswere at carrying capacity before management this 10difference in remaining population could mean 96 more daysbefore the population in Site 3 returned to carrying capacity (iepotentially more time with fewer pigs causing damage on thelandscape) This can be an important distinction when theobjective is to reduce damage which requires deciding howmuch control is cost-effective

We also demonstrated how repeated flights can have asubstantial impact on the population Naiumlve removal ratesbased on the number of animals seen but not killed (due topigs taking shelter in dense cover or no-fly zones)were very high(88 in this study and 94 in Davis et al 2017) Our removalmodel demonstrates that the removal rate per flight was ~27(based on an eight-hour flight and our estimated hourly removalrate of 4) sincemany pigs go undetected Our results show that

702 Wildlife Research A J Davis et al

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

objective All study areas are on private land andwere conductedat the request of the landowner

We used aerial gunning from helicopters to remove wild pigsfrom each study area Aerial gunning was conducted by theUnited States Department of Agriculture (USDA) Animal andPlant Health Inspection Service (APHIS) Wildlife Services(WS) which provides wildlife control assistance to landowners based on the authority of the Animal Damage ControlAct of 1931 (7 USCA x 8351 ndash 8354) and in compliance with theNational Environmental Policy Act (42 USC x 4321) Theobjective was to remove (kill) all pigs encountered the searchstrategy used was based on the expert opinion of the pilotand gunner team and included searching the entire study areaFlights over Site 1 were conducted in April 2016 flights overSite 2 were conducted in June 2016 and flights over Site 3 wereconducted in July 2016 Three flights per site were conducted onsubsequent days We had two pilot and gunner teams (A and B)Team A had 10ndash12 years of experience conducting aerialgunning and Team B had 5ndash7 years of experience conductingaerial gunning In Site 1 TeamAflew on the first day then TeamBon the secondday and thenTeamAagainon the third dayThiswas the same for Site 3 but the order was switched for Site 2Flights were conducted with minimal winds (considerably lessthan the visual flight rules requirements of 26 kn) and noprecipitation Total flight times were generally between 6 and8 h per day excluding time on the ground for refueling Twelve-gauge shotgunswere usedwith double-aught buckshotCarewastaken to ensure all shot animals were killed (animal was seen tocollapse and was often double tapped ndash two shots fired in rapidsuccession at the same target and shot animals were flown backover to verify theywere down beforemoving on) The number ofanimals removed and the hours in the helicopter were recordedfor each flight

We estimated abundance from the aerial gunning datausing a removal model that simultaneously estimatesabundance and removal rates from counts of removedindividuals (Farnsworth et al 2002 Royle and Dorazio 2006)We implemented the removal model in a hierarchical Bayesianframework that accounted for variation in removal effort (Daviset al 2016) Removal models require multiple reduction eventswithin a time period in which there is demographic closure(no births deaths immigration and emigration) Removalestimators are advantageous for monitoring abundance ofinvasive species because they directly use data from controlevents (Zippin 1958 Farnsworth et al 2002 Chee and Wintle2010) for estimation of abundance pre-control From theestimates of abundance pre-control and numbers ofindividuals removed the proportion by which the populationis reduced due to control can be derived We also incorporatedcovariates on removal rate specifically personnel and habitat(see SupplementaryMaterial 1 for model details) We comparedthe removal rates for the two pilot and gunner teams andexamined the relationship with the proportion of forest cover(as determined from Homer et al 2015) We estimated thepopulation remaining after each flight by subtracting thenumber removed by flight from the estimate of initialabundance We then derived the proportion of the populationremoved by each flight by dividing the number removed by eachflight by the estimated population size pre-flight

To examine the long-term impact that reductions had onpopulations we used a standard logistic population growthmodel (Sibly and Hone 2002 Otto and Day 2007) to predictabundance at 1 3 and 5 years following a reduction event Weused estimates of population growth rate and density dependencefrom previous studies to provide the boundaries for ourmodelling (Caley 1993 Bieber and Ruf 2005 Timmons et al2012 Mellish et al 2014) Previous studies on wild pigpopulations have estimated growth rates ranging from 085 to163 (Bieber and Ruf 2005) with the majority of growth-rateestimates between 115 and 135with amean of 124 (Bieber andRuf 2005 Timmons et al 2012 Mellish et al 2014 Moretti2014) We examined a range of growth rates (l) from 1 to 2 toprovide a broader picture of population growth rates that may beexhibited

Management programs for wild pigs are often aimedat reducing damage caused by wild pigs Damage costs canrange from quantifiable assessments of lost crops infrastructurerepairs and livestock feed loss to more abstract costsassociated with disease risks landscape rooting and predationor competition of native imperilled or game species Damagecosts are a combination of all of these factors however there areareas with little or no current damage costs but with the potentialfor great damage if wild pigs are able to establish a population inan area Therefore damage reduction can be in current costs aswell as the prevention of future costs

The relationship between wild pig population densities anddamage incurred bywild pigs is not well known but studies havesuggested that the damagendashdensity relationship for species ingeneral is not necessarily linear (Hone 1995West and Parkhurst2002) To demonstrate the importance of knowing the shape ofthe damagendashdensity relationship for wild pigs we show theimpacts of damage under three potential relationships (Fig 1)linear (damage and density are linearly related) saturating (even

100

Saturating Linear Exponential

100

Population

D

amag

e

80

80

60

60

40

40

20

20

0

0

Fig 1 Theoretical relationships between the percentage of a populationremaining and the percentage of total damage remaining

698 Wildlife Research A J Davis et al

at low levels of density the damage is high) and exponential(only at high levels of density is the damage high)We comparedpercentage damage reduction after 1 3 and 5 years by varyinglevels of percentage population reduction (ranging from 0to 100 reduction) and growth rate (l= 1ndash2) under the threedamage relationships

The damage was not explicitly estimated directly at the sitesin our study Therefore we examined hypothetical damage coststo examine the costndashbenefits of different levels of managementeffort Damage can be to crops pastureland or infrastructureIn our study areas the damage was on pastureland The costsmay depend on the type of damage so we examined a range ofvalues based on average values for different land types peracre and different numbers of acres The range is flexible($0ndash$150 000) to account for different types of damage anddifferent sizes of properties We used the range of damage coststo compare the relative damage costs and management costsof wild pigs across a range of percentage population reductions(0 to 100) Estimated costs of helicopter removal work wasbased on the hourly rate charged for the reduction work weconducted (USD625 per hour)

Results

We removed 324 147 and 362 wild pigs from Sites 1 2 and 3respectively across all flights conducted The total numberremoved by flight and the time spent in the helicopter areshown by site and flight in Table 1 We estimated initialabundance at Sites 1ndash3 to be 413 (95 credible intervals ndash

CI 297 ndash 481) 268 (95 CI 200ndash318) and 545 (95 CI387ndash646) respectively (Fig 2) The average proportion of thepopulation removed after the first flight was 31 (se = 6)cumulatively after the second flight was 56 (se = 15) andcumulatively after the third flight was 67 (se = 12)

Removal rates differed based on the pilot and gunner team(Fig 3A) In our example Team A had a lower hourly removalrate (003 95 CI 002 004) than Team B (005 95 CI 004006 Fig 3A) We found removal rate decreased as percentagecover increased (b= ndash055 95 CI ndash079 ndash027 Fig 3B)

Growth rates impacted the linear damage relationshipmore substantially than either the saturating or the exponentialrelationships (Fig 4) At slower growth rates there was still

Table 1 Raw data from pilot study removing wild pigs from threestudy sites in Texas from aerial gunning work by site

Habitat Pilot andgunner team

Flight No wildpigs removed

Hours

Site 1 Team A 1 157 8Team B 2 140 8Team A 3 27 67

Site 2 Team B 1 71 75Team A 2 43 75Team B 3 33 8

Site 3 Team A 1 161 7Team B 2 125 69Team A 3 76 59

700

600

500

400

300

200

100

38

72

78

2643

55

30

52

66

0Site 1 Site 2 Site 3

Pass 3Pass 2Pass 1Initial

Abu

ndan

ce

Fig 2 Abundance estimate (with 95 confidence intervals for the initialabundance) by site for the initial population size and the subsequent size aftereach flight The proportion of the population removed is shown in text byflight above the abundance estimate

004

003

005

(a) (b)

Rem

ovea

l rat

e

002

Team A Team B 0 20

Site 1

Site 3

Site 2

40 60 80 100

Cover

004

002

000

Fig 3 (A) Boxplot of the posterior distributions of hourly removal rates by pilot and gunner team The greyboxes represent the interquartile range the dotted lines show the 95 credible intervals and the circles representthe total rangeof the posterior distributions (B)Relationship of hourly removal rate and thepercentage cover theshaded region represent the 95 credible intervals Points shown represent the percentage forest cover by site

Cost effectiveness of damage management Wildlife Research 699

damage reduction for all densityndashdamage relationships after5 yearswhen 40or less of the populationwas removed (Fig 4)

The hours of aerial gunning required grew exponentially asthe proportion of the population removed increased (Fig 5A)When the hourly removal rate was greater than 01 the effortrequired to remove up to 90 of a population was reasonable(lt20 h offlight time Fig 5B) If the hourly removal ratewas 004(as was the average in our study) the effort needed to removemore than 90 of the population was more than 45 h (Fig 5A)

The relationship between the reduction in damage costsand removal costs is an important factor in evaluating thecost-effectiveness of management actions For a saturatingdamagendashdensity relationship the population needed to bereduced substantially (80 or more) for the damage reductioncosts to outweigh the removal costs (Fig 6A) The total costsfor the linear relationship and the exponential relationshiponly showed high overall costs when the initial damage costswere high and the population reduction was low (Fig 6B C)

We compared the overall costs with the initial damage costs todetermine if the resulting removal effort resulted in an economicgain or loss over doing nothing For the saturating relationshipwe found that damage costs must be greater than USD50 000and more than 60 of the population must be removed to see aneconomic gain over baseline damage costs (Fig 6D) If thedamage relationships were either linear or exponential therewas generally a cost saving to removing at least some portion ofthe population for almost all levels of damage (Fig 6E F)

Discussion

For other ungulate species the damagendashdensity relationship hasbeen estimated to be high even at low densities ndash similar to thesaturating relationship we presented (Gill 1992 Hone 2002West and Parkhurst 2002) Under the saturating relationshipwe found that the proportion of the population that needs to beremoved in order to have a substantial effect on damage can be

20

18

16

14

100

80

60

40

20

0

12

10

20

18

16

Gro

wth

rat

e

Afte

r 3

Year

sA

fter

5 Ye

ars

Afte

r 1

Year

14

12

10

20

18

16

14

12

1020 40 60 80 20 40 60

Population

Saturating Linear

Damage-density relationshipsExponential

80 20 40 60 80

100

o

f Dam

age80

60

40

20

0

100

80

60

40

20

0

Fig 4 Percentage of total damage (as calculated without any wild pig reduction) by the percentage of the initialwild pig population (eg 20 population corresponds to an 80 reduction in wild pigs) and potential values ofwild pig growth rates The percentage of total damage is shown cumulatively for three time frames (1 3 and 5 yearspost reduction event shown in rows) and for three different potential damagendashdensity relationship (saturating linearand exponential shown in columns)

700 Wildlife Research A J Davis et al

greater than 80 If a saturating relationship is representativeof the damagendashdensity relationship for wild pigs our resultssuggest that it may be more cost-effective to conduct repeatedremoval events within a short time frame than to spread thesame effort over a longer period or a larger area especially at

high-damage sites In contrast if the densityndashdamagerelationship is linear or exponential substantial damagereduction can occur by reducing the population by 50 oronly 20 respectively both of which are substantiallydifferent levels of management effort Thus empirical work

100

(a) (b)

100

005

Population reduced

010

Rem

oval

rat

e

015

020

80

20

20

0

0

60

6040 100

Effo

rt (

eg

hou

rs in

a h

elic

opte

r)

25 +

20

15

10

5

80200 6040

Effo

rt r

equi

red

(in h

ours

)

Fig 5 (A) The effort required is plotted against the percentage of the population removed based on a removalrate for one hour of effort of 004whichwas estimated fromour data (B) Reduction effort (hours in a helicopter)necessary to remove varying proportions of the population (x-axis) by the removal rate given one hour of effort(y-axis)

150 000

(a) (b) (c)

(d ) (e) (f )

100 000

50 000

Tota

l cos

tsC

ost s

avin

gs

0

150 000

100 000

50 000

0

150 000

ndash150 000

100 000

ndash100 000

50 000

ndash50 000

0

150 000

100 000

50 000

000 02 04 06 08 10 00 02 04 06

Proportion population removed08 10 00 02 04 06 08 10

Saturated Linear Exponential

Fig 6 Estimated total costs (damage costs plusmanagement costs top row) andcost savingsover doingnothing (bottom row) for a rangeofinitial damage costs (y-axes) and proportions of the population removed (x-axes) under three damagendashdensity relationship scenariossaturated (AandD) linear (BandE) andexponential (CandF) The total costs are shownranging from lowcosts (blue) tohighcosts (red)Thecost savings are shown inwhitewhen the costs are similar to the initial damage costs blue for cost savings and red for additional costs over theinitial damage costs

Cost effectiveness of damage management Wildlife Research 701

to determine the shape of this relationship is needed formanagersto estimate the most cost-effective actions for a particular levelof damage Once this relationship is better understood we cancompare actual management costs to damage costs

The damagendashdensity relationship may not be constant acrossspace and time and management efforts are often conductedunder the assumption that the damage reduction is greater atdifferent times of year For instance the relationship betweendamage and density may be saturating or linear at some times ofthe year but exponential at others Managers may time removalevents to occur immediately before planting or harvesting ofspecific crops such as peanuts because these crops are onlytargeted by pigs at those times of the year In these cases doingonly two removal events ndash one right before planting and oneright before harvest ndash may be the most cost-effective damagereduction strategy compared with conducting several flightsin the middle of the growing season Additionally theremoval rate is known to vary at different times of the yeardue to weather and seasonal deciduous canopy cover Thereforein addition to the fluctuating potential to reduce damagethe cost-effectiveness may change throughout the year due tochanges in the costs of removing individuals These are standardconsiderations that managers account for when planningremoval work and are necessary to account for in models formanagement effectiveness

We have provided a simplified version of how adamagendashdensity relationship may look and shown howimportant that relationship can be in connecting managementto the objective of damage reduction Single removal events canbe beneficial for many reasons even if the cost calculations donot suggest the reduction in damage is greater than the cost ofremovals Aerial efforts may be employed to access sites thatare not feasible to access from ground efforts (eg trappingor ground shooting) or to target areas that are particularlyvulnerable to damage (eg predation of livestock growthstage of crops) Additionally the damagendashdensity relationshipmay be more complex because different age classes or thedifferent sexes might have a different impact on the damageincurred While young pigs may be more influential onpopulation growth a single adult boar can cause a substantialamount of damage to the landscape Thus it may be mosteffective to use a removal strategy that targets the subset ofthe population that has the largest effect on the underlyingobjective (eg long-term population reduction or immediatedamage reduction)

Reducing wild pig populations has the potential benefit ofreducing damage costs however the reduction events are notwithout costs themselves As a larger portion of the populationis targeted for reduction the costs of reduction increaseexponentially (our results and Choquenot et al 1999) Thecosts to implement reductions in populations can be comparedwith the potential cost savings of damage reductions whichmay help justify expenses for removal work Examining thesecost trade-offs up front can be considerably beneficial forbudget planning In particular population projections can helpjustify the need for large up-front contributions to managementefforts because these large initial investments may result inconsiderably lower long-term costs In contrast consistentsmaller investments in removal work over time may have the

samemanagement costs as one initial intensive reduction effortbut the long-term damage costs may be considerably higheras these may be simply lsquomowing the grassrsquo and not pushing apopulation towards elimination

The damagendashdensity relationship heavily influences howmuch population reduction is necessary to induce a costsavings Generally if the relationship is linear or exponentialdoing almost any removal results in a cost savings unless thereare little or no damage costs Additionally it is important toconsider that reducing future damage costs by reducingthe spread of the species to new areas and preventing damagefromoccurring may result in future savings Site 1 is an examplewhere the primary objective was to prevent the spread of wildpigs on the border of their range in Texas to prevent futuredamage in neighbouring areas Thus even if current damagecosts are low anticipated future damage costs may warrantan immediate expenditure Our results on the relationshipbetween damage costs and population density could be usedeither for planningmanagement against current damage levels oranticipated damage levels However the cost savings we showare short-term savings not accounting for potential effects ofmanagement on suppressing geographic spread which couldlead to additional cost savings that we have not incorporatedpresenting a challenge for future work There is considerablevalue in examining these potential future costs that may resultif wild pigs are allowed to spread to new areas An actualaccounting of the theoretical damage to a new area may makea strong case to focus removal efforts in areas with little currentdamage but that are on the edge of an invasion front instead ofin areas with high current damage since the long-term damagemaybe considerably higher if the invasion is allowed to continue

Repeated reduction events allowed us to estimate theabundance of wild pigs in each of our study areas whilesimultaneously reducing their numbers In our study areas wereduced the populations by an average of 31 after one flightand an average of 67 by the completion of three flights Ifwe had not conducted multiple removals and thus were not ableto estimate abundance we would only have known that we hadremoved 157 71 and 161 individuals in the respective sites(after one flight) In this case we may have expected that wehad impacted Sites 1 and 3 similarly because they had a similarnumber of animals removed when in fact we had removedalmost 40 of the population in Site 1 and only 30 of thepopulation in Site 3 If we were to assume that both populationswere at carrying capacity before management this 10difference in remaining population could mean 96 more daysbefore the population in Site 3 returned to carrying capacity (iepotentially more time with fewer pigs causing damage on thelandscape) This can be an important distinction when theobjective is to reduce damage which requires deciding howmuch control is cost-effective

We also demonstrated how repeated flights can have asubstantial impact on the population Naiumlve removal ratesbased on the number of animals seen but not killed (due topigs taking shelter in dense cover or no-fly zones)were very high(88 in this study and 94 in Davis et al 2017) Our removalmodel demonstrates that the removal rate per flight was ~27(based on an eight-hour flight and our estimated hourly removalrate of 4) sincemany pigs go undetected Our results show that

702 Wildlife Research A J Davis et al

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

at low levels of density the damage is high) and exponential(only at high levels of density is the damage high)We comparedpercentage damage reduction after 1 3 and 5 years by varyinglevels of percentage population reduction (ranging from 0to 100 reduction) and growth rate (l= 1ndash2) under the threedamage relationships

The damage was not explicitly estimated directly at the sitesin our study Therefore we examined hypothetical damage coststo examine the costndashbenefits of different levels of managementeffort Damage can be to crops pastureland or infrastructureIn our study areas the damage was on pastureland The costsmay depend on the type of damage so we examined a range ofvalues based on average values for different land types peracre and different numbers of acres The range is flexible($0ndash$150 000) to account for different types of damage anddifferent sizes of properties We used the range of damage coststo compare the relative damage costs and management costsof wild pigs across a range of percentage population reductions(0 to 100) Estimated costs of helicopter removal work wasbased on the hourly rate charged for the reduction work weconducted (USD625 per hour)

Results

We removed 324 147 and 362 wild pigs from Sites 1 2 and 3respectively across all flights conducted The total numberremoved by flight and the time spent in the helicopter areshown by site and flight in Table 1 We estimated initialabundance at Sites 1ndash3 to be 413 (95 credible intervals ndash

CI 297 ndash 481) 268 (95 CI 200ndash318) and 545 (95 CI387ndash646) respectively (Fig 2) The average proportion of thepopulation removed after the first flight was 31 (se = 6)cumulatively after the second flight was 56 (se = 15) andcumulatively after the third flight was 67 (se = 12)

Removal rates differed based on the pilot and gunner team(Fig 3A) In our example Team A had a lower hourly removalrate (003 95 CI 002 004) than Team B (005 95 CI 004006 Fig 3A) We found removal rate decreased as percentagecover increased (b= ndash055 95 CI ndash079 ndash027 Fig 3B)

Growth rates impacted the linear damage relationshipmore substantially than either the saturating or the exponentialrelationships (Fig 4) At slower growth rates there was still

Table 1 Raw data from pilot study removing wild pigs from threestudy sites in Texas from aerial gunning work by site

Habitat Pilot andgunner team

Flight No wildpigs removed

Hours

Site 1 Team A 1 157 8Team B 2 140 8Team A 3 27 67

Site 2 Team B 1 71 75Team A 2 43 75Team B 3 33 8

Site 3 Team A 1 161 7Team B 2 125 69Team A 3 76 59

700

600

500

400

300

200

100

38

72

78

2643

55

30

52

66

0Site 1 Site 2 Site 3

Pass 3Pass 2Pass 1Initial

Abu

ndan

ce

Fig 2 Abundance estimate (with 95 confidence intervals for the initialabundance) by site for the initial population size and the subsequent size aftereach flight The proportion of the population removed is shown in text byflight above the abundance estimate

004

003

005

(a) (b)

Rem

ovea

l rat

e

002

Team A Team B 0 20

Site 1

Site 3

Site 2

40 60 80 100

Cover

004

002

000

Fig 3 (A) Boxplot of the posterior distributions of hourly removal rates by pilot and gunner team The greyboxes represent the interquartile range the dotted lines show the 95 credible intervals and the circles representthe total rangeof the posterior distributions (B)Relationship of hourly removal rate and thepercentage cover theshaded region represent the 95 credible intervals Points shown represent the percentage forest cover by site

Cost effectiveness of damage management Wildlife Research 699

damage reduction for all densityndashdamage relationships after5 yearswhen 40or less of the populationwas removed (Fig 4)

The hours of aerial gunning required grew exponentially asthe proportion of the population removed increased (Fig 5A)When the hourly removal rate was greater than 01 the effortrequired to remove up to 90 of a population was reasonable(lt20 h offlight time Fig 5B) If the hourly removal ratewas 004(as was the average in our study) the effort needed to removemore than 90 of the population was more than 45 h (Fig 5A)

The relationship between the reduction in damage costsand removal costs is an important factor in evaluating thecost-effectiveness of management actions For a saturatingdamagendashdensity relationship the population needed to bereduced substantially (80 or more) for the damage reductioncosts to outweigh the removal costs (Fig 6A) The total costsfor the linear relationship and the exponential relationshiponly showed high overall costs when the initial damage costswere high and the population reduction was low (Fig 6B C)

We compared the overall costs with the initial damage costs todetermine if the resulting removal effort resulted in an economicgain or loss over doing nothing For the saturating relationshipwe found that damage costs must be greater than USD50 000and more than 60 of the population must be removed to see aneconomic gain over baseline damage costs (Fig 6D) If thedamage relationships were either linear or exponential therewas generally a cost saving to removing at least some portion ofthe population for almost all levels of damage (Fig 6E F)

Discussion

For other ungulate species the damagendashdensity relationship hasbeen estimated to be high even at low densities ndash similar to thesaturating relationship we presented (Gill 1992 Hone 2002West and Parkhurst 2002) Under the saturating relationshipwe found that the proportion of the population that needs to beremoved in order to have a substantial effect on damage can be

20

18

16

14

100

80

60

40

20

0

12

10

20

18

16

Gro

wth

rat

e

Afte

r 3

Year

sA

fter

5 Ye

ars

Afte

r 1

Year

14

12

10

20

18

16

14

12

1020 40 60 80 20 40 60

Population

Saturating Linear

Damage-density relationshipsExponential

80 20 40 60 80

100

o

f Dam

age80

60

40

20

0

100

80

60

40

20

0

Fig 4 Percentage of total damage (as calculated without any wild pig reduction) by the percentage of the initialwild pig population (eg 20 population corresponds to an 80 reduction in wild pigs) and potential values ofwild pig growth rates The percentage of total damage is shown cumulatively for three time frames (1 3 and 5 yearspost reduction event shown in rows) and for three different potential damagendashdensity relationship (saturating linearand exponential shown in columns)

700 Wildlife Research A J Davis et al

greater than 80 If a saturating relationship is representativeof the damagendashdensity relationship for wild pigs our resultssuggest that it may be more cost-effective to conduct repeatedremoval events within a short time frame than to spread thesame effort over a longer period or a larger area especially at

high-damage sites In contrast if the densityndashdamagerelationship is linear or exponential substantial damagereduction can occur by reducing the population by 50 oronly 20 respectively both of which are substantiallydifferent levels of management effort Thus empirical work

100

(a) (b)

100

005

Population reduced

010

Rem

oval

rat

e

015

020

80

20

20

0

0

60

6040 100

Effo

rt (

eg

hou

rs in

a h

elic

opte

r)

25 +

20

15

10

5

80200 6040

Effo

rt r

equi

red

(in h

ours

)

Fig 5 (A) The effort required is plotted against the percentage of the population removed based on a removalrate for one hour of effort of 004whichwas estimated fromour data (B) Reduction effort (hours in a helicopter)necessary to remove varying proportions of the population (x-axis) by the removal rate given one hour of effort(y-axis)

150 000

(a) (b) (c)

(d ) (e) (f )

100 000

50 000

Tota

l cos

tsC

ost s

avin

gs

0

150 000

100 000

50 000

0

150 000

ndash150 000

100 000

ndash100 000

50 000

ndash50 000

0

150 000

100 000

50 000

000 02 04 06 08 10 00 02 04 06

Proportion population removed08 10 00 02 04 06 08 10

Saturated Linear Exponential

Fig 6 Estimated total costs (damage costs plusmanagement costs top row) andcost savingsover doingnothing (bottom row) for a rangeofinitial damage costs (y-axes) and proportions of the population removed (x-axes) under three damagendashdensity relationship scenariossaturated (AandD) linear (BandE) andexponential (CandF) The total costs are shownranging from lowcosts (blue) tohighcosts (red)Thecost savings are shown inwhitewhen the costs are similar to the initial damage costs blue for cost savings and red for additional costs over theinitial damage costs

Cost effectiveness of damage management Wildlife Research 701

to determine the shape of this relationship is needed formanagersto estimate the most cost-effective actions for a particular levelof damage Once this relationship is better understood we cancompare actual management costs to damage costs

The damagendashdensity relationship may not be constant acrossspace and time and management efforts are often conductedunder the assumption that the damage reduction is greater atdifferent times of year For instance the relationship betweendamage and density may be saturating or linear at some times ofthe year but exponential at others Managers may time removalevents to occur immediately before planting or harvesting ofspecific crops such as peanuts because these crops are onlytargeted by pigs at those times of the year In these cases doingonly two removal events ndash one right before planting and oneright before harvest ndash may be the most cost-effective damagereduction strategy compared with conducting several flightsin the middle of the growing season Additionally theremoval rate is known to vary at different times of the yeardue to weather and seasonal deciduous canopy cover Thereforein addition to the fluctuating potential to reduce damagethe cost-effectiveness may change throughout the year due tochanges in the costs of removing individuals These are standardconsiderations that managers account for when planningremoval work and are necessary to account for in models formanagement effectiveness

We have provided a simplified version of how adamagendashdensity relationship may look and shown howimportant that relationship can be in connecting managementto the objective of damage reduction Single removal events canbe beneficial for many reasons even if the cost calculations donot suggest the reduction in damage is greater than the cost ofremovals Aerial efforts may be employed to access sites thatare not feasible to access from ground efforts (eg trappingor ground shooting) or to target areas that are particularlyvulnerable to damage (eg predation of livestock growthstage of crops) Additionally the damagendashdensity relationshipmay be more complex because different age classes or thedifferent sexes might have a different impact on the damageincurred While young pigs may be more influential onpopulation growth a single adult boar can cause a substantialamount of damage to the landscape Thus it may be mosteffective to use a removal strategy that targets the subset ofthe population that has the largest effect on the underlyingobjective (eg long-term population reduction or immediatedamage reduction)

Reducing wild pig populations has the potential benefit ofreducing damage costs however the reduction events are notwithout costs themselves As a larger portion of the populationis targeted for reduction the costs of reduction increaseexponentially (our results and Choquenot et al 1999) Thecosts to implement reductions in populations can be comparedwith the potential cost savings of damage reductions whichmay help justify expenses for removal work Examining thesecost trade-offs up front can be considerably beneficial forbudget planning In particular population projections can helpjustify the need for large up-front contributions to managementefforts because these large initial investments may result inconsiderably lower long-term costs In contrast consistentsmaller investments in removal work over time may have the

samemanagement costs as one initial intensive reduction effortbut the long-term damage costs may be considerably higheras these may be simply lsquomowing the grassrsquo and not pushing apopulation towards elimination

The damagendashdensity relationship heavily influences howmuch population reduction is necessary to induce a costsavings Generally if the relationship is linear or exponentialdoing almost any removal results in a cost savings unless thereare little or no damage costs Additionally it is important toconsider that reducing future damage costs by reducingthe spread of the species to new areas and preventing damagefromoccurring may result in future savings Site 1 is an examplewhere the primary objective was to prevent the spread of wildpigs on the border of their range in Texas to prevent futuredamage in neighbouring areas Thus even if current damagecosts are low anticipated future damage costs may warrantan immediate expenditure Our results on the relationshipbetween damage costs and population density could be usedeither for planningmanagement against current damage levels oranticipated damage levels However the cost savings we showare short-term savings not accounting for potential effects ofmanagement on suppressing geographic spread which couldlead to additional cost savings that we have not incorporatedpresenting a challenge for future work There is considerablevalue in examining these potential future costs that may resultif wild pigs are allowed to spread to new areas An actualaccounting of the theoretical damage to a new area may makea strong case to focus removal efforts in areas with little currentdamage but that are on the edge of an invasion front instead ofin areas with high current damage since the long-term damagemaybe considerably higher if the invasion is allowed to continue

Repeated reduction events allowed us to estimate theabundance of wild pigs in each of our study areas whilesimultaneously reducing their numbers In our study areas wereduced the populations by an average of 31 after one flightand an average of 67 by the completion of three flights Ifwe had not conducted multiple removals and thus were not ableto estimate abundance we would only have known that we hadremoved 157 71 and 161 individuals in the respective sites(after one flight) In this case we may have expected that wehad impacted Sites 1 and 3 similarly because they had a similarnumber of animals removed when in fact we had removedalmost 40 of the population in Site 1 and only 30 of thepopulation in Site 3 If we were to assume that both populationswere at carrying capacity before management this 10difference in remaining population could mean 96 more daysbefore the population in Site 3 returned to carrying capacity (iepotentially more time with fewer pigs causing damage on thelandscape) This can be an important distinction when theobjective is to reduce damage which requires deciding howmuch control is cost-effective

We also demonstrated how repeated flights can have asubstantial impact on the population Naiumlve removal ratesbased on the number of animals seen but not killed (due topigs taking shelter in dense cover or no-fly zones)were very high(88 in this study and 94 in Davis et al 2017) Our removalmodel demonstrates that the removal rate per flight was ~27(based on an eight-hour flight and our estimated hourly removalrate of 4) sincemany pigs go undetected Our results show that

702 Wildlife Research A J Davis et al

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

damage reduction for all densityndashdamage relationships after5 yearswhen 40or less of the populationwas removed (Fig 4)

The hours of aerial gunning required grew exponentially asthe proportion of the population removed increased (Fig 5A)When the hourly removal rate was greater than 01 the effortrequired to remove up to 90 of a population was reasonable(lt20 h offlight time Fig 5B) If the hourly removal ratewas 004(as was the average in our study) the effort needed to removemore than 90 of the population was more than 45 h (Fig 5A)

The relationship between the reduction in damage costsand removal costs is an important factor in evaluating thecost-effectiveness of management actions For a saturatingdamagendashdensity relationship the population needed to bereduced substantially (80 or more) for the damage reductioncosts to outweigh the removal costs (Fig 6A) The total costsfor the linear relationship and the exponential relationshiponly showed high overall costs when the initial damage costswere high and the population reduction was low (Fig 6B C)

We compared the overall costs with the initial damage costs todetermine if the resulting removal effort resulted in an economicgain or loss over doing nothing For the saturating relationshipwe found that damage costs must be greater than USD50 000and more than 60 of the population must be removed to see aneconomic gain over baseline damage costs (Fig 6D) If thedamage relationships were either linear or exponential therewas generally a cost saving to removing at least some portion ofthe population for almost all levels of damage (Fig 6E F)

Discussion

For other ungulate species the damagendashdensity relationship hasbeen estimated to be high even at low densities ndash similar to thesaturating relationship we presented (Gill 1992 Hone 2002West and Parkhurst 2002) Under the saturating relationshipwe found that the proportion of the population that needs to beremoved in order to have a substantial effect on damage can be

20

18

16

14

100

80

60

40

20

0

12

10

20

18

16

Gro

wth

rat

e

Afte

r 3

Year

sA

fter

5 Ye

ars

Afte

r 1

Year

14

12

10

20

18

16

14

12

1020 40 60 80 20 40 60

Population

Saturating Linear

Damage-density relationshipsExponential

80 20 40 60 80

100

o

f Dam

age80

60

40

20

0

100

80

60

40

20

0

Fig 4 Percentage of total damage (as calculated without any wild pig reduction) by the percentage of the initialwild pig population (eg 20 population corresponds to an 80 reduction in wild pigs) and potential values ofwild pig growth rates The percentage of total damage is shown cumulatively for three time frames (1 3 and 5 yearspost reduction event shown in rows) and for three different potential damagendashdensity relationship (saturating linearand exponential shown in columns)

700 Wildlife Research A J Davis et al

greater than 80 If a saturating relationship is representativeof the damagendashdensity relationship for wild pigs our resultssuggest that it may be more cost-effective to conduct repeatedremoval events within a short time frame than to spread thesame effort over a longer period or a larger area especially at

high-damage sites In contrast if the densityndashdamagerelationship is linear or exponential substantial damagereduction can occur by reducing the population by 50 oronly 20 respectively both of which are substantiallydifferent levels of management effort Thus empirical work

100

(a) (b)

100

005

Population reduced

010

Rem

oval

rat

e

015

020

80

20

20

0

0

60

6040 100

Effo

rt (

eg

hou

rs in

a h

elic

opte

r)

25 +

20

15

10

5

80200 6040

Effo

rt r

equi

red

(in h

ours

)

Fig 5 (A) The effort required is plotted against the percentage of the population removed based on a removalrate for one hour of effort of 004whichwas estimated fromour data (B) Reduction effort (hours in a helicopter)necessary to remove varying proportions of the population (x-axis) by the removal rate given one hour of effort(y-axis)

150 000

(a) (b) (c)

(d ) (e) (f )

100 000

50 000

Tota

l cos

tsC

ost s

avin

gs

0

150 000

100 000

50 000

0

150 000

ndash150 000

100 000

ndash100 000

50 000

ndash50 000

0

150 000

100 000

50 000

000 02 04 06 08 10 00 02 04 06

Proportion population removed08 10 00 02 04 06 08 10

Saturated Linear Exponential

Fig 6 Estimated total costs (damage costs plusmanagement costs top row) andcost savingsover doingnothing (bottom row) for a rangeofinitial damage costs (y-axes) and proportions of the population removed (x-axes) under three damagendashdensity relationship scenariossaturated (AandD) linear (BandE) andexponential (CandF) The total costs are shownranging from lowcosts (blue) tohighcosts (red)Thecost savings are shown inwhitewhen the costs are similar to the initial damage costs blue for cost savings and red for additional costs over theinitial damage costs

Cost effectiveness of damage management Wildlife Research 701

to determine the shape of this relationship is needed formanagersto estimate the most cost-effective actions for a particular levelof damage Once this relationship is better understood we cancompare actual management costs to damage costs

The damagendashdensity relationship may not be constant acrossspace and time and management efforts are often conductedunder the assumption that the damage reduction is greater atdifferent times of year For instance the relationship betweendamage and density may be saturating or linear at some times ofthe year but exponential at others Managers may time removalevents to occur immediately before planting or harvesting ofspecific crops such as peanuts because these crops are onlytargeted by pigs at those times of the year In these cases doingonly two removal events ndash one right before planting and oneright before harvest ndash may be the most cost-effective damagereduction strategy compared with conducting several flightsin the middle of the growing season Additionally theremoval rate is known to vary at different times of the yeardue to weather and seasonal deciduous canopy cover Thereforein addition to the fluctuating potential to reduce damagethe cost-effectiveness may change throughout the year due tochanges in the costs of removing individuals These are standardconsiderations that managers account for when planningremoval work and are necessary to account for in models formanagement effectiveness

We have provided a simplified version of how adamagendashdensity relationship may look and shown howimportant that relationship can be in connecting managementto the objective of damage reduction Single removal events canbe beneficial for many reasons even if the cost calculations donot suggest the reduction in damage is greater than the cost ofremovals Aerial efforts may be employed to access sites thatare not feasible to access from ground efforts (eg trappingor ground shooting) or to target areas that are particularlyvulnerable to damage (eg predation of livestock growthstage of crops) Additionally the damagendashdensity relationshipmay be more complex because different age classes or thedifferent sexes might have a different impact on the damageincurred While young pigs may be more influential onpopulation growth a single adult boar can cause a substantialamount of damage to the landscape Thus it may be mosteffective to use a removal strategy that targets the subset ofthe population that has the largest effect on the underlyingobjective (eg long-term population reduction or immediatedamage reduction)

Reducing wild pig populations has the potential benefit ofreducing damage costs however the reduction events are notwithout costs themselves As a larger portion of the populationis targeted for reduction the costs of reduction increaseexponentially (our results and Choquenot et al 1999) Thecosts to implement reductions in populations can be comparedwith the potential cost savings of damage reductions whichmay help justify expenses for removal work Examining thesecost trade-offs up front can be considerably beneficial forbudget planning In particular population projections can helpjustify the need for large up-front contributions to managementefforts because these large initial investments may result inconsiderably lower long-term costs In contrast consistentsmaller investments in removal work over time may have the

samemanagement costs as one initial intensive reduction effortbut the long-term damage costs may be considerably higheras these may be simply lsquomowing the grassrsquo and not pushing apopulation towards elimination

The damagendashdensity relationship heavily influences howmuch population reduction is necessary to induce a costsavings Generally if the relationship is linear or exponentialdoing almost any removal results in a cost savings unless thereare little or no damage costs Additionally it is important toconsider that reducing future damage costs by reducingthe spread of the species to new areas and preventing damagefromoccurring may result in future savings Site 1 is an examplewhere the primary objective was to prevent the spread of wildpigs on the border of their range in Texas to prevent futuredamage in neighbouring areas Thus even if current damagecosts are low anticipated future damage costs may warrantan immediate expenditure Our results on the relationshipbetween damage costs and population density could be usedeither for planningmanagement against current damage levels oranticipated damage levels However the cost savings we showare short-term savings not accounting for potential effects ofmanagement on suppressing geographic spread which couldlead to additional cost savings that we have not incorporatedpresenting a challenge for future work There is considerablevalue in examining these potential future costs that may resultif wild pigs are allowed to spread to new areas An actualaccounting of the theoretical damage to a new area may makea strong case to focus removal efforts in areas with little currentdamage but that are on the edge of an invasion front instead ofin areas with high current damage since the long-term damagemaybe considerably higher if the invasion is allowed to continue

Repeated reduction events allowed us to estimate theabundance of wild pigs in each of our study areas whilesimultaneously reducing their numbers In our study areas wereduced the populations by an average of 31 after one flightand an average of 67 by the completion of three flights Ifwe had not conducted multiple removals and thus were not ableto estimate abundance we would only have known that we hadremoved 157 71 and 161 individuals in the respective sites(after one flight) In this case we may have expected that wehad impacted Sites 1 and 3 similarly because they had a similarnumber of animals removed when in fact we had removedalmost 40 of the population in Site 1 and only 30 of thepopulation in Site 3 If we were to assume that both populationswere at carrying capacity before management this 10difference in remaining population could mean 96 more daysbefore the population in Site 3 returned to carrying capacity (iepotentially more time with fewer pigs causing damage on thelandscape) This can be an important distinction when theobjective is to reduce damage which requires deciding howmuch control is cost-effective

We also demonstrated how repeated flights can have asubstantial impact on the population Naiumlve removal ratesbased on the number of animals seen but not killed (due topigs taking shelter in dense cover or no-fly zones)were very high(88 in this study and 94 in Davis et al 2017) Our removalmodel demonstrates that the removal rate per flight was ~27(based on an eight-hour flight and our estimated hourly removalrate of 4) sincemany pigs go undetected Our results show that

702 Wildlife Research A J Davis et al

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

greater than 80 If a saturating relationship is representativeof the damagendashdensity relationship for wild pigs our resultssuggest that it may be more cost-effective to conduct repeatedremoval events within a short time frame than to spread thesame effort over a longer period or a larger area especially at

high-damage sites In contrast if the densityndashdamagerelationship is linear or exponential substantial damagereduction can occur by reducing the population by 50 oronly 20 respectively both of which are substantiallydifferent levels of management effort Thus empirical work

100

(a) (b)

100

005

Population reduced

010

Rem

oval

rat

e

015

020

80

20

20

0

0

60

6040 100

Effo

rt (

eg

hou

rs in

a h

elic

opte

r)

25 +

20

15

10

5

80200 6040

Effo

rt r

equi

red

(in h

ours

)

Fig 5 (A) The effort required is plotted against the percentage of the population removed based on a removalrate for one hour of effort of 004whichwas estimated fromour data (B) Reduction effort (hours in a helicopter)necessary to remove varying proportions of the population (x-axis) by the removal rate given one hour of effort(y-axis)

150 000

(a) (b) (c)

(d ) (e) (f )

100 000

50 000

Tota

l cos

tsC

ost s

avin

gs

0

150 000

100 000

50 000

0

150 000

ndash150 000

100 000

ndash100 000

50 000

ndash50 000

0

150 000

100 000

50 000

000 02 04 06 08 10 00 02 04 06

Proportion population removed08 10 00 02 04 06 08 10

Saturated Linear Exponential

Fig 6 Estimated total costs (damage costs plusmanagement costs top row) andcost savingsover doingnothing (bottom row) for a rangeofinitial damage costs (y-axes) and proportions of the population removed (x-axes) under three damagendashdensity relationship scenariossaturated (AandD) linear (BandE) andexponential (CandF) The total costs are shownranging from lowcosts (blue) tohighcosts (red)Thecost savings are shown inwhitewhen the costs are similar to the initial damage costs blue for cost savings and red for additional costs over theinitial damage costs

Cost effectiveness of damage management Wildlife Research 701

to determine the shape of this relationship is needed formanagersto estimate the most cost-effective actions for a particular levelof damage Once this relationship is better understood we cancompare actual management costs to damage costs

The damagendashdensity relationship may not be constant acrossspace and time and management efforts are often conductedunder the assumption that the damage reduction is greater atdifferent times of year For instance the relationship betweendamage and density may be saturating or linear at some times ofthe year but exponential at others Managers may time removalevents to occur immediately before planting or harvesting ofspecific crops such as peanuts because these crops are onlytargeted by pigs at those times of the year In these cases doingonly two removal events ndash one right before planting and oneright before harvest ndash may be the most cost-effective damagereduction strategy compared with conducting several flightsin the middle of the growing season Additionally theremoval rate is known to vary at different times of the yeardue to weather and seasonal deciduous canopy cover Thereforein addition to the fluctuating potential to reduce damagethe cost-effectiveness may change throughout the year due tochanges in the costs of removing individuals These are standardconsiderations that managers account for when planningremoval work and are necessary to account for in models formanagement effectiveness

We have provided a simplified version of how adamagendashdensity relationship may look and shown howimportant that relationship can be in connecting managementto the objective of damage reduction Single removal events canbe beneficial for many reasons even if the cost calculations donot suggest the reduction in damage is greater than the cost ofremovals Aerial efforts may be employed to access sites thatare not feasible to access from ground efforts (eg trappingor ground shooting) or to target areas that are particularlyvulnerable to damage (eg predation of livestock growthstage of crops) Additionally the damagendashdensity relationshipmay be more complex because different age classes or thedifferent sexes might have a different impact on the damageincurred While young pigs may be more influential onpopulation growth a single adult boar can cause a substantialamount of damage to the landscape Thus it may be mosteffective to use a removal strategy that targets the subset ofthe population that has the largest effect on the underlyingobjective (eg long-term population reduction or immediatedamage reduction)

Reducing wild pig populations has the potential benefit ofreducing damage costs however the reduction events are notwithout costs themselves As a larger portion of the populationis targeted for reduction the costs of reduction increaseexponentially (our results and Choquenot et al 1999) Thecosts to implement reductions in populations can be comparedwith the potential cost savings of damage reductions whichmay help justify expenses for removal work Examining thesecost trade-offs up front can be considerably beneficial forbudget planning In particular population projections can helpjustify the need for large up-front contributions to managementefforts because these large initial investments may result inconsiderably lower long-term costs In contrast consistentsmaller investments in removal work over time may have the

samemanagement costs as one initial intensive reduction effortbut the long-term damage costs may be considerably higheras these may be simply lsquomowing the grassrsquo and not pushing apopulation towards elimination

The damagendashdensity relationship heavily influences howmuch population reduction is necessary to induce a costsavings Generally if the relationship is linear or exponentialdoing almost any removal results in a cost savings unless thereare little or no damage costs Additionally it is important toconsider that reducing future damage costs by reducingthe spread of the species to new areas and preventing damagefromoccurring may result in future savings Site 1 is an examplewhere the primary objective was to prevent the spread of wildpigs on the border of their range in Texas to prevent futuredamage in neighbouring areas Thus even if current damagecosts are low anticipated future damage costs may warrantan immediate expenditure Our results on the relationshipbetween damage costs and population density could be usedeither for planningmanagement against current damage levels oranticipated damage levels However the cost savings we showare short-term savings not accounting for potential effects ofmanagement on suppressing geographic spread which couldlead to additional cost savings that we have not incorporatedpresenting a challenge for future work There is considerablevalue in examining these potential future costs that may resultif wild pigs are allowed to spread to new areas An actualaccounting of the theoretical damage to a new area may makea strong case to focus removal efforts in areas with little currentdamage but that are on the edge of an invasion front instead ofin areas with high current damage since the long-term damagemaybe considerably higher if the invasion is allowed to continue

Repeated reduction events allowed us to estimate theabundance of wild pigs in each of our study areas whilesimultaneously reducing their numbers In our study areas wereduced the populations by an average of 31 after one flightand an average of 67 by the completion of three flights Ifwe had not conducted multiple removals and thus were not ableto estimate abundance we would only have known that we hadremoved 157 71 and 161 individuals in the respective sites(after one flight) In this case we may have expected that wehad impacted Sites 1 and 3 similarly because they had a similarnumber of animals removed when in fact we had removedalmost 40 of the population in Site 1 and only 30 of thepopulation in Site 3 If we were to assume that both populationswere at carrying capacity before management this 10difference in remaining population could mean 96 more daysbefore the population in Site 3 returned to carrying capacity (iepotentially more time with fewer pigs causing damage on thelandscape) This can be an important distinction when theobjective is to reduce damage which requires deciding howmuch control is cost-effective

We also demonstrated how repeated flights can have asubstantial impact on the population Naiumlve removal ratesbased on the number of animals seen but not killed (due topigs taking shelter in dense cover or no-fly zones)were very high(88 in this study and 94 in Davis et al 2017) Our removalmodel demonstrates that the removal rate per flight was ~27(based on an eight-hour flight and our estimated hourly removalrate of 4) sincemany pigs go undetected Our results show that

702 Wildlife Research A J Davis et al

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

to determine the shape of this relationship is needed formanagersto estimate the most cost-effective actions for a particular levelof damage Once this relationship is better understood we cancompare actual management costs to damage costs

The damagendashdensity relationship may not be constant acrossspace and time and management efforts are often conductedunder the assumption that the damage reduction is greater atdifferent times of year For instance the relationship betweendamage and density may be saturating or linear at some times ofthe year but exponential at others Managers may time removalevents to occur immediately before planting or harvesting ofspecific crops such as peanuts because these crops are onlytargeted by pigs at those times of the year In these cases doingonly two removal events ndash one right before planting and oneright before harvest ndash may be the most cost-effective damagereduction strategy compared with conducting several flightsin the middle of the growing season Additionally theremoval rate is known to vary at different times of the yeardue to weather and seasonal deciduous canopy cover Thereforein addition to the fluctuating potential to reduce damagethe cost-effectiveness may change throughout the year due tochanges in the costs of removing individuals These are standardconsiderations that managers account for when planningremoval work and are necessary to account for in models formanagement effectiveness

We have provided a simplified version of how adamagendashdensity relationship may look and shown howimportant that relationship can be in connecting managementto the objective of damage reduction Single removal events canbe beneficial for many reasons even if the cost calculations donot suggest the reduction in damage is greater than the cost ofremovals Aerial efforts may be employed to access sites thatare not feasible to access from ground efforts (eg trappingor ground shooting) or to target areas that are particularlyvulnerable to damage (eg predation of livestock growthstage of crops) Additionally the damagendashdensity relationshipmay be more complex because different age classes or thedifferent sexes might have a different impact on the damageincurred While young pigs may be more influential onpopulation growth a single adult boar can cause a substantialamount of damage to the landscape Thus it may be mosteffective to use a removal strategy that targets the subset ofthe population that has the largest effect on the underlyingobjective (eg long-term population reduction or immediatedamage reduction)

Reducing wild pig populations has the potential benefit ofreducing damage costs however the reduction events are notwithout costs themselves As a larger portion of the populationis targeted for reduction the costs of reduction increaseexponentially (our results and Choquenot et al 1999) Thecosts to implement reductions in populations can be comparedwith the potential cost savings of damage reductions whichmay help justify expenses for removal work Examining thesecost trade-offs up front can be considerably beneficial forbudget planning In particular population projections can helpjustify the need for large up-front contributions to managementefforts because these large initial investments may result inconsiderably lower long-term costs In contrast consistentsmaller investments in removal work over time may have the

samemanagement costs as one initial intensive reduction effortbut the long-term damage costs may be considerably higheras these may be simply lsquomowing the grassrsquo and not pushing apopulation towards elimination

The damagendashdensity relationship heavily influences howmuch population reduction is necessary to induce a costsavings Generally if the relationship is linear or exponentialdoing almost any removal results in a cost savings unless thereare little or no damage costs Additionally it is important toconsider that reducing future damage costs by reducingthe spread of the species to new areas and preventing damagefromoccurring may result in future savings Site 1 is an examplewhere the primary objective was to prevent the spread of wildpigs on the border of their range in Texas to prevent futuredamage in neighbouring areas Thus even if current damagecosts are low anticipated future damage costs may warrantan immediate expenditure Our results on the relationshipbetween damage costs and population density could be usedeither for planningmanagement against current damage levels oranticipated damage levels However the cost savings we showare short-term savings not accounting for potential effects ofmanagement on suppressing geographic spread which couldlead to additional cost savings that we have not incorporatedpresenting a challenge for future work There is considerablevalue in examining these potential future costs that may resultif wild pigs are allowed to spread to new areas An actualaccounting of the theoretical damage to a new area may makea strong case to focus removal efforts in areas with little currentdamage but that are on the edge of an invasion front instead ofin areas with high current damage since the long-term damagemaybe considerably higher if the invasion is allowed to continue

Repeated reduction events allowed us to estimate theabundance of wild pigs in each of our study areas whilesimultaneously reducing their numbers In our study areas wereduced the populations by an average of 31 after one flightand an average of 67 by the completion of three flights Ifwe had not conducted multiple removals and thus were not ableto estimate abundance we would only have known that we hadremoved 157 71 and 161 individuals in the respective sites(after one flight) In this case we may have expected that wehad impacted Sites 1 and 3 similarly because they had a similarnumber of animals removed when in fact we had removedalmost 40 of the population in Site 1 and only 30 of thepopulation in Site 3 If we were to assume that both populationswere at carrying capacity before management this 10difference in remaining population could mean 96 more daysbefore the population in Site 3 returned to carrying capacity (iepotentially more time with fewer pigs causing damage on thelandscape) This can be an important distinction when theobjective is to reduce damage which requires deciding howmuch control is cost-effective

We also demonstrated how repeated flights can have asubstantial impact on the population Naiumlve removal ratesbased on the number of animals seen but not killed (due topigs taking shelter in dense cover or no-fly zones)were very high(88 in this study and 94 in Davis et al 2017) Our removalmodel demonstrates that the removal rate per flight was ~27(based on an eight-hour flight and our estimated hourly removalrate of 4) sincemany pigs go undetected Our results show that

702 Wildlife Research A J Davis et al

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

additional flights lead to observing and thus removingmore pigsand can produce costs savings (Fig 6) suggesting that multipleflights over the same space can be cost-effective especially whendamage reduction is the objective Since we demonstratedhow removal rates can vary by personnel and habitat (Fig 3)the cost-effectiveness will vary under different personnel andhabitat types However it is always important to recognise thatas resources are limited flying multiple times in one area willprevent havingmoreflights over awider area Thereforewemaysee considerable damage reduction in one site but this couldresult in no damage reduction in other sites The short-termdamage costs may be equivalent at a broader spatial scale whenconducting intensive removals in one area or less intensiveremovals on a broader area The long-term benefits may varyand examinations of these trade-offs is another important factorto consider for efficient management planning

Another advantage to knowing the proportion of thepopulation reduced by management is that the longer-term(in our case up to 5 years) effects of population reduction canbe estimated when changes in abundance are known Our resultsshow that longer-term impact to a population was more affectedby the proportion of the population removed compared withthe growth rate except when the growth rate was high (gt18)indicating the importance of removing a larger percentage of thepopulation to optimise management efficacy on a property or inan area

Wildlife management is always conducted under constrainedresources while attempting to address multiple managementobjectives (eg conserving native and game species reducinginvasive or overabundant species protecting natural oragricultural resources) while satisfying the diverse interestsof stakeholders (eg land owners government agencies and aplethora of publics) Therefore managing an overabundantpopulation like wild pigs needs not only to be consideredin terms of the most cost-effective strategy but also in whatdecisions maintain or improve stakeholder opinions andsupport In some cases it may not be cost-effective norlogistically possible to visit a property multiple timeshowever the benefit in terms of fostering good relationshipsmay be invaluable The management strategies that ultimatelyare most beneficial for the landscape as a whole will be a balanceof strategies that foster good relationships with stakeholders andones that reduce the populations most effectively

Management implications

We showed how cost-effectiveness of wild pig damagemanagement by aerial gunning can be evaluated usingremoval data and population models Despite the commonbelief that it is not cost-effective to conduct repeated flightsover the same space within a short time frame our resultsshowed that repeat flights can increase cost-effectiveness ndash

especially when damage levels are high or the damagerelationship is saturating However if damage levels are lowor if the damagendashdensity relationship is exponential conductingone flight might be the most cost-effective approach Our studywas focussed on three sites in Texas where aerial gunning forwild pigs is a common management technique In other states orin locationswith generally lower densities therewould be amuch

different costndashbenefit calculation However the rational formonitoring during management and considering costs ofdamage and removal still hold In cases where resources forconducting damage management are limited and locations areprioritised by landowner demand our approach to estimatingprogram-wide cost-effectiveness could be applied to datafrom a subset of areas to make inference in the broadersystem Therefore targeted checks on population efficiencyand cost-effectiveness may be a practical approach whilegaining important information

Conflicts of interest

The authors declare no conflicts of interest

Acknowledgements

Thanks to theWildlife Services personnel (ThomasTaylor LanceNixon andCole Taylor) who conducted aerial work for this project We would like tothank two anonymous reviewers for helpful comments on this manuscriptAny use of trade firm or product names is for descriptive purposes only anddoes not imply endorsement by the US government

This research did not receive any specific funding

References

Barrios-Garcia M N and Ballari S A (2012) Impact of wild boar (Susscrofa) in its introduced and native range a reviewBiological Invasions14 2283ndash2300 doi101007s10530-012-0229-6

Baxter P W J Sabo J L Wilcox C McCarthy M A and PossinghamH P (2008) Cost-effective suppression and eradication of invasivepredators Conservation Biology 22 89ndash98 doi101111j1523-1739200700850x

BengsenA J Leung LK P Lapidge S J andGordon I J (2011)Usinga general index approach to analyze camera-trap abundance indices TheJournal of Wildlife Management 75 1222ndash1227 doi101002jwmg132

Bengsen A J Gentle M N Mitchell J L Pearson H E and SaundersG R (2014) Impacts and management of wild pigs Sus scrofa inAustralia Mammal Review 44 135ndash147 doi101111mam12011

Bevins S N Pedersen K Lutman M W Gidlewski T and DelibertoT J (2014) Consequences associated with the recent range expansionof nonnative feral swine Bioscience 64 291ndash299 doi101093bioscibiu015

Bieber C and Ruf T (2005) Population dynamics in wild boar Susscrofa ecology elasticity of growth rate and implications for themanagement of pulsed resource consumers Journal of AppliedEcology 42 1203ndash1213 doi101111j1365-2664200501094x

Bleier N Lehoczki R Uacutejvaacutery D Szemethy L and Csaacutenyi S (2012)Relationships between wild ungulates density and crop damage inHungary Acta Theriologica 57 351ndash359 doi101007s13364-012-0082-0

Buckland S T Anderson D R Burnham K P Laake J L BorchersD L and Thomas L (2001) lsquoIntroduction to Distance Samplingestimating abundance of biological populationsrsquo (Oxford UniversityPress Oxford UK)

Caley P (1993) Population dynamics of feral pigs (Sus scrofa) in a tropicalriverine habitat complex Wildlife Research 20 625ndash636 doi101071WR9930625

Caley P (1994) Factors affecting the success rate of traps for catching feralpigs in a tropical habitat Wildlife Research 21 287ndash292 doi101071WR9940287

Campbell T A and Long D B (2009) Feral swine damage and damagemanagement in forested ecosystems Forest Ecology and Management257 2319ndash2326 doi101016jforeco200903036

Cost effectiveness of damage management Wildlife Research 703

Campbell T A LongD B and Leland B R (2010) Feral swine behaviorrelative to aerial gunning in southern Texas The Journal of WildlifeManagement 74 337ndash341 doi1021932009-131

Chee Y E and Wintle B A (2010) Linking modelling monitoringand management an integrated approach to controlling overabundantwildlife Journal of Applied Ecology 47 1169ndash1178 doi101111j1365-2664201001877x

Choquenot D Kilgour R J and Lukins B S (1993) An evaluationof feral pig trapping Wildlife Research 20 15ndash21 doi101071WR9930015

Choquenot D Hone J and Saunders G (1999) Using aspects ofpredatorndashprey theory to evaluate helicopter shooting for feral pigcontrol Wildlife Research 26 251ndash261 doi101071WR98006

Conover M R (2001) lsquoResolving Humanndashwildlife Conflicts the Scienceof Wildlife Damage Managementrsquo (CRC Press Boca Raton FL)

Cowled B D Elsworth P and Lapidge S J (2008) Additional toxins forferal pig (Sus scrofa) control identifying and testing Achillesrsquo heelsWildlife Research 35 651ndash662 doi101071WR07072

Davis A J Hooten M B Miller R S Farnsworth M L Lewis JMoxcey M and Pepin K M (2016) Inferring invasive speciesabundance using removal data from management actions EcologicalApplications 26 2339ndash2346 doi101002eap1383

Davis A J Leland B Bodenchuk M VerCauteren K C and PepinKM (2017) Estimating population density for disease risk assessmentthe importance of understanding the area of influence of traps usingwild pigs as an example Preventive Veterinary Medicine 141 33ndash37doi101016jprevetmed201704004

DeMaso S J Guthery F S Spears G S and Rice SM (1992)Morningcovey calls as an index of northern bobwhite density Wildlife SocietyBulletin 20 94ndash101

Diefenbach D R Palmer W L and Shope W K (1997) Attitudes ofPennsylvania sportsmen towards managing white-tailed deer to protectthe ecological integrity of forestsWildlife Society Bulletin 25 244ndash251

Fackler P L andHaight R G (2014)Monitoring as a partially observabledecision problem Resource and Energy Economics 37 226ndash241doi101016jreseneeco201312005

Fagerstone K A and Clay W H (1997) Overview of USDA animaldamage control efforts to manage overabundant deer Wildlife SocietyBulletin 25 413ndash417

Farnsworth G L Pollock K H Nichols J D Simons T R Hines J ESauer J R and Brawn J (2002) A removal model for estimatingdetection probabilities from point-count surveys The Auk 119 414ndash425doi1016420004-8038(2002)119[0414ARMFED]20CO2

Gill R M A (1992) A review of damage by mammals in north temperateforests 1 Deer Forestry International Journal of Forestry Research 65145ndash169

Gove N E Skalski J R Zager P and Townsend R L (2002) Statisticalmodels for population reconstruction using age-at-harvest data TheJournal of Wildlife Management 66 310ndash320 doi1023073803163

Hauser C E Pople A R and Possingham H P (2006) Should managedpopulations be monitored every year Ecological Applications 16807ndash819 doi1018901051-0761(2006)016[0807SMPBME]20CO2

Homer C G Dewitz J A Yang L Jin S Danielson P Xian GCoulston J Herold N D Wickham J D and Megown K (2015)Completion of the 2011 National Land Cover Database for theconterminous United States-Representing a decade of land coverchange information Photogrammetric Engineering amp Remote Sensing81 345ndash354

Hone J (1995) Spatial and temporal aspects of vertebrate pest damagewith emphasis on feral pigs Journal of Applied Ecology 32 311ndash319doi1023072405098

Hone J (1996) Analysis of vertebrate pest research In lsquoProceedings ofthe Seventeenth Vertebrate Pest Conferencersquo (Eds RM Timm amp ACCrabb) pp 13ndash17 (University of California Davis CA)

Hone J (2002) Feral pigs in Namadgi National Park Australia dynamicsimpacts and management Biological Conservation 105 231ndash242doi101016S0006-3207(01)00185-9

Linnell J D Broseth H Odden J and Nilsen E B (2010)Sustainably harvesting a large carnivore Development of Eurasianlynx populations in Norway during 160 years of shifting policyEnvironmental Management 45 1142ndash1154 doi101007s00267-010-9455-9

Lyons J E Runge M C Laskowski H P and Kendall W L (2008)Monitoring in the context of structured decision-making and adaptivemanagement The Journal of Wildlife Management 72 1683ndash1692doi1021932008-141

MasseiGCowanD Coats JGladwell F Lane J andMiller L (2008)Effect of the GnRH vaccine GonaCon on the fertility physiology andbehaviour of wild boar Wildlife Research 35 540ndash547 doi101071WR07132

Mellish J M Sumrall A Campbell T A Collier B A Neill W HHigginbotham B and Lopez R R (2014) Simulating potentialpopulation growth of wild pig Sus scrofa in Texas SoutheasternNaturalist (Steuben ME) 13 367ndash376 doi1016560580130217

MorettiM (1995)Biometric data andgrowth rates of amountainpopulationof wild boar (Sus scrofa L) Ticino Switzerland IBEX ndash Journal ofMountain Ecology 3 56ndash59

Naughton-Treves L (1998) Predicting patterns of crop damage by wildlifearound Kibale National Park Uganda Conservation Biology 12156ndash168 doi101046j1523-1739199896346x

Otto S P and Day T (2007) lsquoA Biologistrsquos Guide to MathematicalModeling in Ecology and Evolutionrsquo Vol 13 (Princeton UniversityPress Princeton IL)

Parkes J P Ramsey D S L Macdonald N Walker K McKnight SCohen B S andMorrison S A (2010) Rapid eradication of feral pigs(Sus scrofa) fromSantaCruz Island CaliforniaBiological Conservation143 634ndash641 doi101016jbiocon200911028

Pimentel D (2007) Environmental and economic costs of vertebratespecies invasions into the United States In lsquoManaging VertebrateInvasive Species Proceedings of an International Symposiumrsquo (EdsG WWitmer W C Pitt K A Fagerston) pp 13ndash17 (US Departmentof Agriculture Animal and Plant Health Inspection Service WildlifeServices National Wildlife Research Center Fort Collins CO)

Royle J A and Dorazio R M (2006) Hierarchical models of animalabundance and occurrence Journal of Agricultural Biological ampEnvironmental Statistics 11 249ndash263 doi101198108571106X129153

Rupp S P Ballard W B and Wallace M C (2000) A nationwideevaluation of deer hunter harvest survey techniques Wildlife SocietyBulletin 28 570ndash578

Seward N W Ver Cauteren K C Witmer G W and Engeman R M(2004) Feral swine impacts on agriculture and the environment SheepampGoat Research Journal 12 34ndash40

Shrnhur A J Levy O and Dvir D (1997) Mapping the dimensions ofproject success Project Management Journal 28 5ndash13

Sibly R M and Hone J (2002) Population growth rate and itsdeterminants an overview Philosophical Transactions of the RoyalSociety of London Series B Biological Sciences 357 1153ndash1170doi101098rstb20021117

Slade N A and Blair S M (2000) An empirical test of using countsof individuals captured as indices of population size Journal ofMammalogy 81 1035ndash1045 doi1016441545-1542(2000)081lt1035AETOUCgt20CO2

SnowN P JarzynaMA andVerCauteren K C (2017) Interpreting andpredicting the spread of invasive wild pigs Journal of Applied Ecology54 2022ndash2032 doi1011111365-266412866

St Clair K Dunton E and Giudice J (2012) A comparison of modelsusing removal effort to estimate animal abundance Journal of AppliedStatistics 40 527ndash545 doi101080026647632012748016

704 Wildlife Research A J Davis et al

Timmons J B Mellish J Higginbotham B Griffin J Lopez R RSumrall A Cathey J C and Skow K (2012) Feral Hog PopulationGrowth Density andHarvest in Texas Texas AampMAgriLife ExtensionService Report (Texas AampM University College Station TX)

VerCauteren K C Anderson C W van Deelen T R Drake D WalterW D Vantassel S M and Hygnstrom S E (2011) Regulatedcommercial harvest to manage overabundant white-tailed deer anidea to consider Wildlife Society Bulletin 35 185ndash194 doi101002wsb36

Waithman J D Sweitzer R A Vuren DV Drew J D Brinkhaus A JGardner I A and Boyce W M (1999) Range expansion population

sizes andmanagement ofwild pigs in CaliforniaThe Journal ofWildlifeManagement 63 298ndash308 doi1023073802513

West B C and Parkhurst J A (2002) Interactions betweenDeer DamageDeer Density and Stakeholder Attitudes in Virginia Wildlife SocietyBulletin 30 139ndash147

Williams B L Holtfreter R W Ditchkoff S S and Grand J B (2011)Trap style influenceswild pig behavior and trapping successThe Journalof Wildlife Management 75 432ndash436 doi101002jwmg64

Zippin C (1958) The removal method of population estimation TheJournal of Wildlife Management 22 82ndash90 doi1023073797301

Cost effectiveness of damage management Wildlife Research 705

wwwpublishcsiroaujournalswr