Brown bear damage: patterns and hotspots in Croatia - Cambridge University Press

Page created by Marshall Scott
 
CONTINUE READING
Brown bear damage: patterns and hotspots in Croatia - Cambridge University Press
Brown bear damage: patterns and hotspots in Croatia
                      DÁRIO HIPÓLITO, SLAVEN RELJIĆ, LUÍS MIGUEL ROSALINO, SETH M. WILSON
                                                            C A R L O S F O N S E C A and Đ U R O H U B E R

           Abstract Human–bear conflicts resulting from livestock                                 Introduction
           depredation and crop use are a common threat to the brown
           bear Ursus arctos throughout its range. Understanding these
           conflicts requires the recording and categorization of in-
           cidents, assessment of their geographical distribution and
                                                                                                  N      egative interactions between people and wildlife, often
                                                                                                         referred to as human–wildlife conflict, are one of the
                                                                                                  primary challenges to large carnivore conservation (Treves
           frequency, and documentation of the financial costs and                                & Karanth, ), particularly in Europe where human po-
           the presence of any preventative measures. Damage com-                                 pulations and activities often overlap with the ranges of
           pensation schemes can help mitigate conflicts and, in                                  large carnivores. Such negative interactions are often related
           some cases, improve acceptance of bears. This study                                    to wild animals’ use of biological resources that are pro-
           aims to elucidate the major factors determining the pat-                               duced (e.g. crops, livestock, beehives) or exploited (e.g.
           terns of damage caused by bears, examine the effective-                                game) by humans (Kruuk, ). The evaluation of such
           ness of preventative measures in reducing such damage,                                 damage, the underlying drivers and the efficacy of damage
           and identify bear damage hotspots in Croatia. Our ana-                                 prevention and compensation frameworks is crucial for
           lysis is based on damage reports provided by hunting or-                               large carnivore conservation (Schwerdtner & Gruber,
           ganizations to the Croatian Ministry of Agriculture                                    ; Rigg et al., ). This is particularly important for
           during –. The highest number of claims were                                    species such as the brown bear Ursus arctos because damage
           made for damage to field crops and orchards. Damage                                    caused by bears tends to attract less public attention than
           to livestock, agricultural crops and beehives resulted in                              that caused by wolves or large felids. As a result, public in-
           the highest total cost to farmers. Damage to beehives                                  stitutions focus less on assessing the impact of such damage
           and to automatic corn feeders for game species incurred                                on rural populations, which limits the efficiency of mitiga-
           the highest cost per damage event. We identified a hotspot                             tion measures (Can et al., ). In this analysis we use the
           for bear damage claims in Croatia, located near Risnjak                                term conflict to refer broadly to incidents caused by brown
           National Park and the border with Slovenia. Damage ap-                                 bears and the term damage to refer specifically to economic
           pears higher in areas that have more villages closer to pro-                           losses.
           tected areas and a greater per cent of forest cover,                                       Brown bears were present historically throughout con-
           indicating a synergistic effect of protected environments                              tinental Europe (Zedrosser et al., ; Trouwborst, ),
           that facilitate bear movements and the presence of                                     but were nearly extirpated from western and southern
           human activities that provide easily accessible food for                               Europe and from many areas in eastern and northern
           bears.                                                                                 Europe before World War II, primarily as a result of defor-
                                                                                                  estation and human persecution (Zedrosser et al., ;
           Keywords Brown bear, Croatia, damage compensation,
                                                                                                  Huber et al., a; Trouwborst, ). However, bear be-
           Dinaric Mountains, human–wildlife conflict, large carni-                               haviour is plastic and bears can adapt to certain levels of dis-
           vores, Ursus arctos, wildlife management                                               turbance. This, along with increasing conservation efforts,
           Supplementary material for this article is available at                                rural abandonment and rewilding of many European re-
           https://doi.org/./S                                              gions, has allowed bears to gradually reoccupy former habi-
                                                                                                  tats (Huber et al., a).
                                                                                                      In Croatia brown bears occur in the Dinaric Mountains,
                                                                                                  from Slovenia to Bosnia and Herzegovina and further
                                                                                                  south-east (Servheen et al., ; Zedrosser et al., ),
           DÁRIO HIPÓLITO* (Corresponding author), SLAVEN RELJIĆ and ĐURO HUBER                   with an estimated population of c. , (Kocijan &
           Biology Department, Faculty of Veterinary Medicine, University of Zagreb,              Huber, ; Majić et al., ; Knott et al., ). Brown
           Heinzelova, Zagreb, Croatia. E-mail dhipolito@ua.pt
                                                                                                  bear management is challenging because of the species’ bio-
           LUÍS MIGUEL ROSALINO and CARLOS FONSECA Departamento de Biologia &
           CESAM, Universidade de Aveiro, Aveiro, Portugal
                                                                                                  logical and ecological characteristics (large body, long gesta-
                                                                                                  tion period and opportunistic feeding strategy) and its
           SETH M. WILSON Northern Rockies Conservation Cooperative, Jackson, USA
                                                                                                  history of conflicts with people. In Croatia the majority of
           *Also at: Departamento de Biologia & CESAM, Universidade de Aveiro, Aveiro,
           Portugal                                                                               human–bear conflicts result from livestock depredation
           Received  July . Revision requested  September .
                                                                                                  and crop use (Nyhus et al., ; Majić et al., ;
           Accepted  January . First published online  September .                    Bautista et al., ). The most common strategies to

           Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
Brown bear damage: patterns and hotspots in Croatia - Cambridge University Press
512         D. Hipólito et al.

             minimize these conflicts include prevention tools and com-                             organization responsible for managing the area fills out a re-
             pensation schemes (Treves & Karanth, ). In Croatia the                             port and sends it to the Croatian Ministry of Agriculture,
             latter have been the focal approach, aiming to distribute                              where the information about the type and amount of dam-
             damage costs between the conservation beneficiary (govern-                             age, date and location of the incident, name of the injured
             ment or hunting organizations) and the person or institu-                              party, and presence of any preventative measures are com-
             tion suffering the damage (Fourli, ; Nyhus et al.,                                 piled. We included in our analysis all reported claims that
             ). There are two types of compensation schemes: ex-                                were confirmed as damage caused by bears, including claims
             post compensation, which is paid after damage occurs,                                  that were and were not subsequently compensated for.
             and compensation in advance, which is paid prior to any                                    We analysed the number of damage claims, the cost asso-
             damage occurring and based on an estimation of expected                                ciated with the damage and the frequency of incidents by
             losses (Schwerdtner & Gruber, ). Regardless of the                                 damage type. Damage type was categorized as losses asso-
             type of scheme being used, the priority should be to prevent                           ciated with crops, orchards/vineyards (including domestic
             damage from occurring or to minimize its impact. This re-                              gardens), beehives, and wildlife/game feeders. Corn feeders
             quires an understanding of the types and cost of damage                                are widely used in Croatia to provide extra food for game spe-
             events, their frequency, and the spatial distribution and pat-                         cies such as red deer Cervus elaphus, wild boar Sus scrofa and
             terns of such incidents. (Gunther et al., ; Wilson et al.,                         bears. Additionally, we analysed incidents involving damage
             ; Can et al., ).                                                               to property such as barn doors, vehicles, shooting posts or
                 In Croatia mitigation of human–bear conflict is based pri-                         silage depots, and losses of livestock and domestic animals.
             marily on the ex-post compensatory approach, although bear                             For the latter category, we assessed the number of bear at-
             hunting is also allowed as a strategy to minimize conflicts.                           tacks per type of livestock and domestic animal (poultry, cat-
             Ex-post compensation is managed and paid by local hunting                              tle, red deer, dogs, donkeys, goats, horses, ostriches, pigs,
             organizations, which have an allocated quota for bear hunting                          rabbits and sheep). To test whether there was a positive or
             (Huber et al., a; Majić et al., ; Bišćan et al., ).                        negative trend in the total number of damage events per
             However, when damage occurs in national parks, the govern-                             year we used a linear regression model (Zar, ), and χ
             ment is responsible for paying compensation (Huber et al.,                             tests (Zar, ) to assess if the number of damage events (ob-
             a). Damage will usually be fully compensated only if                               served values) differed between damage types. We gathered
             the person or institution suffering it has previously implemen-                        data on the intensity of livestock depredation by bears across
             ted protection measures (e.g. proper fencing) and behaves re-                          Europe using the estimations of the annual per capita loss
             sponsibly (e.g. by guarding livestock and not planting crops                           presented by Kazcensky (), i.e. the number of livestock
             that entice bears near the forest edge; Huber et al., a,b);                        lost per year and per bear. For Croatia, we used the method-
             but occasionally hunting organizations compensate for dam-                             ology described by Kazcensky () to estimate this metric,
             age in the absence of protection measures by providing the                             allowing us to compare livestock depredation across all
             farmer with an amount of crops equivalent to the damage                                European countries where brown bears occur. The number
             caused by bears. To apply compensation strategies efficiently,                         of bears used for this calculation was the annual estimate of
             it is crucial to review available data, so that damage patterns                        the total bear population in Croatia, as provided by the hunt-
             can be detected and research needs identified (Bruggers                                ing organizations. To identify damage types that have the big-
             et al., ). No such analysis has previously been carried                            gest impact in terms of cost, we defined the following damage
             out in Croatia, and we aim to contribute new insights for a bet-                       categories: Wildlife feeder, Beehive, Orchards, Crops,
             ter understanding of human–bear conflicts. Specifically, our                           Livestock and domestic animals, and Other. We used partial
             objectives are to () analyse brown bear damage in Croatia                             least square regressions (Roy & Roy, ; Carrascal et al.,
             during –, including the type of damage, associated                             ) to determine the influence of these damage categories
             costs, and the factors determining damage patterns, () deter-                         (all binary variables), together with the presence of protective
             mine if preventative tools reduce brown bear damage and as-                            measures on the damage compensation paid. Model signifi-
             sociated costs, and () identify bear damage hotspots.                                 cance was assessed by a Stone–Geisser Q test, which evalu-
                                                                                                    ates the accuracy of models and their parameters compared
                                                                                                    to observed values through a cross-validation process (Götz
             Methods                                                                                et al., ). We used Q $ . (Cao et al., ) to assign
                                                                                                    significance to the contribution of the predictors, and R to
             Damage claims and costs                                                                assess the proportion of the total variance that was explained
                                                                                                    by the model (Zar, ).
             We analysed all damage reports for – that were                                     We evaluated whether the frequency of damage to do-
             systematically collected and provided annually by hunting                              mestic animals decreased when protective measures were
             organizations to the Croatian Ministry of Agriculture.                                 used, by testing whether the number of damage claims dif-
             When damage occurs and is reported, the hunting                                        fered between properties using protection measures (Cat)

                                                                                        Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
Brown bear damage patterns in Croatia                      513

           and those without them (Cat), using a χ test. All analyses                              The landscape protection status hypothesis (H) tests
           were performed using R .. (R Development Core Team,                                 whether areas where bears were not hunted acted as source
           ), and partial least square regression models were built                           environments (Hansen, ) and were therefore subject to
           using the package plsdepot (Sanchez, ).                                            more frequent damage events. The landscape protection sta-
                                                                                                  tus is the distance in km between the edge of a hunting re-
                                                                                                  serve and the nearest edge of a protected area.
           Damage hotspots                                                                           The bear population characteristic hypothesis (H) ex-
                                                                                                  amines whether more stable populations and constant pres-
           We analysed the spatial distribution of damage claims using
                                                                                                  ence of bears lead to more frequent damage events. Bear
           a geographical information system. Hotspot areas were de-
                                                                                                  occurrence was based on the Croatian Bear Management
           lineated using the Optimized Hot Spot Analysis tool in
                                                                                                  Plan and the LIFE DinAlp Bear project report (i.e. constant
           ArcMap .. (Esri, Redlands, USA). Each damage location,
                                                                                                  vs sporadic presence; Huber et al., b; Skrbinšek et al.,
           defined by its latitude and longitude coordinates, was as-
                                                                                                  ) and the number of bears $  years old that were
           signed to a  ×  km grid cell on a map of Croatia. For
                                                                                                  hunted and killed in hunting reserves was used as a surro-
           each grid cell, we tallied the number of damage events and
                                                                                                  gate for population stability.
           calculated the Getis–Ord Gi* statistic (Songchitruksa &
                                                                                                     The hybrid hypothesis (H) used variables included in
           Zeng, ) by scoring it with a z-score and probability.
                                                                                                  the best-fit models from each of the previous hypotheses
           The resulting z-score was based on the distance between
                                                                                                  where the % confidence interval of their coefficients did
           cells and the number of damage events within them.
                                                                                                  not include zero. This allowed us to detect a positive or
           Thus, cells with high numbers of damage events that were
                                                                                                  negative influence on the damage frequency.
           in close proximity to each other produced high z-scores.
                                                                                                     We first tested for data multicollinearity using Spearman’s
           To be characterized as a statistically significant hotspot, a
                                                                                                  correlation coefficient (ρ; Zar, ), and identified outliers
           cell had to be surrounded by cells with high z-scores and
                                                                                                  by using a χ test (Komsta, ). Variables were considered
           probabilities (Ord & Getis, ). These statistics allowed
                                                                                                  highly correlated if ρ . .. When two covariates were cor-
           us to compare the number of damage events in each cell
                                                                                                  related we excluded those less correlated with the dependent
           and its neighbouring cells to the mean number of damage
                                                                                                  variable.
           events per grid cell across the entire study area. Where
                                                                                                     We built models using all possible combinations of
           any observed difference is larger than expected by chance,
                                                                                                  the covariates for each hypothesis and based our model
           a statistically significant z-score is achieved. We used the
                                                                                                  selection on the Akaike Information Criterion corrected
           Getis–Ord Gi* statistic to identify patterns in positive spatial
                                                                                                  for small samples (AICc; Burnham & Anderson, ).
           clustering. This allowed us to discriminate between cells of
                                                                                                  Models with ΔAICc ,  were considered the most parsimo-
           high and low spatial association (Songchitruksa & Zeng,
                                                                                                  nious (Burnham & Anderson, ).
           ), and to generate a density surface raster layer.
                                                                                                     We compared the AICc of the best models in each hy-
                                                                                                  pothesis and that of the hybrid hypothesis (H) using R (R
           Factors influencing bear damage frequency                                              Development Core Team, ) together with the MuMIn
                                                                                                  package (Barton, ). Model evaluation was performed
           To evaluate which factors influence damage frequency (the                              using the pseudo R and the likelihood ratio test comparing
           number of damage claims per area) in each hunting reserve                              the deviance of the null best model (Dobson, ).
           (the damage compensation management unit in Croatia),
           we used generalized linear models (Zuur et al., ), with
           a Gaussian distribution. We defined four hypotheses that                               Results
           could potentially explain differences in damage frequency
           between different areas.                                                               Damage claims and costs
              The land cover hypothesis (H) assumes that herbaceous
           vegetation is a habitat used by bears (Posillico et al., )                         During – a total of  claims for damage caused
           and examines whether damage occurred more frequently in                                by bears were reported to the Croatian Ministry of
           such areas. Land cover variables were extracted from Corine                            Agriculture, a mean of c.  per year. A total of EUR ,
            land cover maps (CLC, ): per cent of forests, scrub                           was paid in compensation during this period, with a mean
           and/or herbaceous vegetation, artificial surfaces or houses,                           of EUR , per year. There was no trend in the annual
           arable land and permanent crops, heterogeneous agricultur-                             number of claims over this period (Spearman correlation
           al areas (a surrogate for the presence of human activities)                            coefficient = −., P = .). There was a significantly
           and wetlands, water bodies, and the number of villages with-                           higher proportion of claims associated with crop damage to
           in each hunting reserve (see CLC, , for definitions of                             corn, oat and hay fields compared to the other types of dam-
           these categories).                                                                     age (% of all damage claims; χ = .; P , .; Fig. ).

           Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
514         D. Hipólito et al.

                                                                                                                                     FIG. 1 Proportion (%) of different
                                                                                                                                     types of asset damaged by brown
                                                                                                                                     bears Ursus arctos of the total
                                                                                                                                     number of bear-related damage
                                                                                                                                     claims and associated costs in Croatia
                                                                                                                                     (–).

                 A different pattern was evident when we analysed the                               TABLE 1 Influence of assessed variables on compensation paid for
             amount of damage costs. The majority of the total compen-                              damage caused by brown bears Ursus arctos in Croatia during
             sation was paid for damage to beehives, crops and domestic                             –, with their loads and weights on the first component
             animals, including livestock (% each; Fig. ). The number                            of partial least squares regression. A variable’s weight is its contri-
             of damage claims and associated costs from beehives, crops                             bution to the first component of partial least squares regression.
             and domestic animals was significantly higher (χ = .;                            Variable                                           Load                Weight
             P , .) where protective measures were absent. Most live-
                                                                                                    Wildlife feeders                                  0.448                 0.482
             stock damage claims were associated with sheep (%,                                 Beehives                                          0.408                 0.406
             sheep killed in  attacks, mean . per attack) and chickens                          Crops                                             0.113                 0.073
             (%;  chickens killed in  attacks, mean . per attack),                         Livestock & domestic animals                      0.108                 0.103
             with multiple individuals usually killed in an incident. Attacks                       Orchards                                         −0.780                −0.750
             on individual animals involved mostly large mammals such                               Other                                             0.082                 0.058
             as donkeys, horses and red deer (which are also raised as                              Protection measures                              −0.064                −0.141
             farm animals in Croatia), but also ostriches and dogs
             (Supplementary Table ). The number of animals killed per                              with beehives (load = ., weight = .; Table ) and
             attack was typically greater where protective measures had                             wildlife feeders (load = ., weight = .; Table ).
             been implemented compared to those where protective mea-                               Conversely, orchard damage contributed less to damage
             sures had not been implemented (Supplementary Table ).                                costs, with a negative correlation with the total cost of dam-
                 The reports showed that  beehives,  of which were                            age (load = −., weight = −.; Table ).
             unprotected, were damaged in  events. The mean number
             of beehives damaged per attack was . (. without and .
             with protection) and the mean monetary value per damaged                               Damage hotspots
             beehive was EUR .
                 The mean intensity of livestock depredation (annual per                            We identified a hotspot where most damage claims were
             capita loss) during – was . livestock per bear                              clustered, in the north-west of the brown bear range in
             per year. For beehives, the intensity of depredation was                               Croatia, near the border with Slovenia (Fig. ).
             . hives per bear per year. These calculations are based
             on the estimated annual bear population in the country,                                Factors influencing bear damage frequency
             which ranged from  in  to , in .
                 Partial least squares regression analysis showed that                              The values for per cent cover of forests and scrub and/or
             c. % of damage cost variability was explained by the                                 herbaceous vegetation were highly correlated (r = −.,
             type of damage and whether protective measurements                                     P , .) and therefore we excluded the latter from the
             were present. The Stone–Geisser Q test value was ., in-                           generalized linear model analysis, as it was less correlated
             dicating that model estimates were accurate (i.e. . .).                           than the former with the dependent variable. We identified
             The amount of compensation was most influenced by the                                  three outliers, which were also excluded. Five, one and three
             type of damage, with higher values primarily associated                                models were identified as the best models for the land cover,

                                                                                        Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
Brown bear damage patterns in Croatia                      515

           FIG. 2 Brown bear distribution (a) and damage claim locations and hotspots (b) in Croatia for –.

           landscape protection status and bear population character-                             the per cent of forest cover within the hunting reserves had a
           istics hypotheses, respectively (Table ; Supplementary                                positive effect (Table ). The best model’s pseudo R was
           Material ). The hybrid hypothesis models had lower                                    . (Table ) and was better fitted than the null model
           AICc values (.–.), with a minimum ΔAICc of                                   (χ = .; P , .; Table ).
           . for the lowest model for the other hypotheses, and
           thus had the highest support. Of the models built for this
           hypothesis, two were the most parsimonious, having                                     Discussion
           ΔAICc ,  (Table ). However, the ΔAICc values of the
           second-best model were close to the threshold of : the differ-                        To address human–bear conflicts effectively it is imperative
           ence is because the best model has one variable less (see AICc                         to identify the types of damage and the geographical distri-
           formula in Burnham & Anderson, ). Model averaging is                               bution of damage events. This is critical for effective wildlife
           less appropriate in such situations and the most parsimonious                          management, conservation planning and for targeting con-
           model should be selected (Banner & Higgs, ). For the                               flict mitigation measures in areas where negative impacts on
           three variables included in the best models of the hybrid hy-                          people and bears can be minimized.
           pothesis, the % confidence interval did not include zero, so                             The majority (%) of compensation claims related to
           it is possible to determine whether their effect on the depend-                        bears in Croatia were made for damage to crops and orch-
           ent value is positive or negative. Distance to the nearest pro-                        ards. In contrast, Bautista et al. () found that . % of
           tected area had a negative effect on the number of bear                                bear damage events in western Europe involved livestock
           damage events per km, whereas the number of villages and                              losses and destruction of apiaries. This difference is

           Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
516         D. Hipólito et al.

             TABLE 2 Best generalized linear regression models (ΔAICc , ) for each hypothesis formulated to explain variation in bear damage fre-
             quency, ranked by the ΔAICc values. Overall best models are in bold.

                                                                                                     Degrees of                                         Overall           Akaike
             Model                                                                                   freedom               AICc1         ΔAICc2         ΔAICc3            weight
             Hypothesis 1: Land cover
               No. of villages + % Forest cover                                                             4             1281.61          0.00           4.70             0.178
               % Forest cover                                                                               3             1282.63          1.02           5.72             0.107
               No. of villages + % Forest cover + % Artificial surfaces & houses                            5             1282.75          1.14           5.84             0.101
               % Forest cover + % Heterogeneous agricultural areas                                          5             1283.20          1.59           6.29             0.08
               % Arable land & permanent crops + % Forest cover                                             5             1283.26          1.65           6.35             0.078
             Hypothesis 2: Landscape protection status
               Distance to protected areas                                                                  3             1284.60          0.00           7.69             1.000
             Hypothesis 3: Bear population characteristics
               No. of bears $ 8 years old hunted + Type of bear occurrence                                  4             1285.24          0.00           8.33             0.400
               Type of bear occurrence                                                                      3             1285.49          0.25           8.58             0.355
               No. of bears $ 8 years old hunted                                                            3             1286.87          1.63           9.96             0.177
             Hypothesis 4: Hybrid
               Distance to protected areas + No. of villages + % Forest cover                               5             1276.91          0.00           0.004            0.395
               Type of bear occurrence + Distance to protected areas + No. of                               6             1278.89          1.98           1.985            0.146
                 villages + % Forest cover
             
              Akaike Information Criterion, adjusted for small sample size.
             
               Difference from best ranking (lowest AIC) model for the same hypothesis.
             
               Difference from best ranking (lowest AIC) model across all hypotheses.
             
               Best model.
             
               Second-best model.

             TABLE 3 Variables included in overall best model (hybrid hypothesis), and their coefficients (standardized using the partial standard de-
             viations), standard errors, z-values, significance and % CI. Variables for which the % CI does not include zero are in bold.

                                                                                                                               Significance
             Variable                                           Coefficient ± SE                     z-value                   P (.|z|)                          95% CI
             Intercept                                          −0.399 ± 1.764                       −0.226                    0.821                             −3.858–3.060
             Distance to protected areas                        −0.080 ± 0.031                       −2.606                    0.010                             −0.140–−0.020
             No. of villages                                     0.158 ± 0.077                        2.064                    0.041                              0.008–0.309
             % Forest cover                                      0.066 ± 0.025                        2.631                    0.009                              0.017–0.115

             probably a result of higher accessibility and availability of                          damage claims was particularly high, with  recorded
             crops and orchards to bears in Croatia, although no system-                            events. Local communities reported a sharp decrease in
             atic data are currently available regarding the availability of                        beechnut Fagus sylvatica and wild berry abundance in this
             such commodities in Croatia. Compensation costs were                                   year (DH, pers. obs.). This shortage of forest foods may have
             highest for damage to crops, livestock or domestic animals,                            led bears closer to villages in search of alternative food
             and beehives, each accounting for c. % of the total com-                             sources and may have resulted in an increased number of
             pensation payments made (i.e. together, representing %                               damage events. Although the bear population appears to
             of the total payments). There was a high, positive correlation                         be increasing in Croatia (Bautista et al., ), the mean
             between the compensation paid and the damage associated                                number of damage claims per year per bear is one of the
             with beehives and wildlife feeders, with the converse for                              lowest in Europe (Bautista et al., ).
             orchards. Beehives and feeder structures are highly valuable                              The application of effective preventative measures or an
             and damage to them often results in complete economic                                  increase in habitat productivity may be contributing to the
             loss. In orchards, bears may feed in a less destructive man-                           relatively low and stable number of damage claims.
             ner, picking fruit from trees and causing less damage as a                             However, further research is needed to examine the me-
             result, although they may occasionally destroy trees                                   chanisms behind this pattern.
             (Mahmoud et al., ).                                                                   A possible factor could be supplementary feeding during
                 There was no trend in the number of damage events per                              hunting periods (March–May and September–December)
             year during –, although in  the number of                                  in Croatia, which, together with abundant wild forest

                                                                                        Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
Brown bear damage patterns in Croatia                      517

           foods, may keep bears away from human settlements be-                                  in Croatia (Kaczensky, ). However, annual per capita
           cause it reduces their need to seek and exploit anthropogen-                           loss estimates are dependent on the robustness of bear popu-
           ic food sources (Bautista et al., ). However, it is the                            lation estimates. We used the official estimates published by
           converse in Slovenia (Jerina et al., ; Kavčič et al.,                              the Croatia government, but these may not be accurate be-
           ), with an increase in bear densities and bear-related                             cause hunting organizations often deal with compensation
           damage events apparently linked to supplementary feeding.                              claims without reporting them to government institutions;
           Such differences may be linked to specific landscape charac-                           our results should therefore be interpreted cautiously. This
           teristics and species distribution patterns, but the exact rea-                        situation could be related to the compensation method
           sons are difficult to determine. The bear population in                                used in Croatia. Because a majority (. %) of hunters are
           Slovenia is concentrated in only % of the country, reach-                            also farmers, and as a hunting right, land owners are respon-
           ing mean densities of  bears/ km and high natality                               sible for damage management in Croatia, they may be more
           rates in those areas, which are supported by supplementary                             protective of their livestock or domestic animals compared to
           feeding sites (Jerina et al., ; Kavčič et al., ). In                           their counterparts elsewhere in Europe, for example by keep-
           Croatia the bear population is concentrated in % of the                              ing their animals in shelters during the night when bears are
           country and the mean density is lower than in Slovenia (                              more active. Most of the damage associated with livestock is
           bears/ km). The higher bear density in Slovenia results                            a result of predation on sheep, one of the preferred prey of
           in a higher number of bears inhabiting areas near human                                brown bears in the area, which is consistent with data col-
           settlements, and the spatial distribution of forests and settle-                       lected in other European countries (e.g. Slovenia; Kavčič
           ments allows bears to visit feeding sites in forests and                               et al., ). Farmers in Croatia traditionally manage sheep
           human settlements during the same night (Kavčič et al.,                                flocks in open areas, making them easier prey for bears
           ), increasing the likelihood of damage.                                            (Fourli, ; Dečak et al., ; Huber et al., b).
               Our data confirm that agricultural or livestock areas                              Similarly, the high number of chickens killed per attack is
           without protective measures are more likely to be damaged                              a result of the management practice of keeping large num-
           by bears. Most damage to beehives and livestock is avoidable,                          bers of chickens in henhouses.
           considering that effective protection measures exist (Swenson                              Beehive annual per capita loss in Croatia (.) is lower
           et al., ). Electric fences, for example, are effective against                     than in other European countries, e.g. Greece (.;
           bears (Coordination Board for Bear Management in Austria,                              Karamanlidis et al., ). This may be a result of the higher
           ), but many farmers are still reluctant to use them, prob-                         numbers of beehives available per location in the brown
           ably because of the associated cost. Little research has been                          bear range in Greece compared to Croatia and probably ex-
           done in Croatia to test the efficacy and functionality of                              plains a higher number of beehives being damaged in an at-
           damage protection measures, but farmers should be involved                             tack when unprotected (Svečnjak et al., ).
           in developing and implementing preventative measures that                                  Preventative measures are an important tool for mitigat-
           could reduce damage claims. Crop protection can be more                                ing brown bear damage, but some stakeholders do not apply
           costly than the protection of livestock and beehives, particu-                         them. We recommend that compensation should be linked
           larly for large fields, but alternative approaches can be imple-                       to the implementation of damage prevention measures
           mented. For example, it is common practice in Croatia for                              (Coordination Board for Bear Management in Austria,
           hunting ground managers to compensate the owners of da-                                ) to encourage their use and ensure a low level of dam-
           maged cereal crop fields by providing an amount of replace-                            age (and consequently of cost and conflict). Hunter-farmers
           ment cereals equal to the crop loss caused by bears rather                             are more likely to implement preventative measures because
           than through financial compensation.                                                   they are involved with the institutions that pay for bear
               Although the number of damage events affecting livestock                           damage. The amount of damage and number of claims
           or domestic animals was lower when protective measures                                 may decrease when protective measures are implemented,
           were in place, there is a tendency for a greater number of an-                         with a consequent amelioration of human–bear conflict.
           imals to be killed per attack. This could be because confined                              We detected a damage hotspot in the north of Croatia
           domestic animals become more vulnerable when bears de-                                 near Risnjak National Park. This protected area (. km)
           stroy or overcome the protective measures, as their ability to                         is relatively small compared to the size of a typical brown
           escape is limited by the barriers put in place to prevent access.                      bear home range ( and  km for males and females,
               Although Croatia has intermediate levels of livestock                              respectively; Huber & Roth, ). However, the Park is a
           density within the European Union (EU; Eurostat, )                                 refuge for wildlife because of its undisturbed habitat and it
           and bears do cause damage to livestock or domestic animals,                            is probably important for bears because it provides dens
           the country’s annual per capita loss of livestock to bears                             where they can hibernate more safely than in agricultural
           (.) is the lowest among European countries. For example,                            areas (Petram et al., ).
           in Sweden, which has one of the lowest EU livestock densities                              The conflict in this area could be a result of the bears’
           (Eurostat, ), the annual per capita loss value is twice that                       population dynamics in the adjacent protected area. Our

           Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
518         D. Hipólito et al.

             model results indicated that hunting reserves close to pro-                                                  brown bear damage on a continental scale. Journal of Applied
             tected areas had a higher damage frequency, and that                                                         Ecology, , –.
                                                                                                                      B I Š Ć A N , A., B U D O R , I., D O M A Z E T O V I Ć , Z., F O N TA N A P U D I Ć , K.,
             areas with higher forest cover and a higher concentration
                                                                                                                          F R A N C E T I Ć , I., G O S P O Č I Ć , S. et al. () Action Plan for Brown
             of villages were more likely to suffer damage. This may in-                                                  Bear Management in Republic of Croatia for . Ministry of
             dicate a synergistic effect of forests as a protective environ-                                              Agriculture and Ministry of Environmental and Nature Protection.
             ment where bears can move freely without being detected                                                      Zagreb, Croatia. [In Croatian]
             (Posillico et al., ), and the presence of human activity,                                            B R U G G E R S , R.L., O W E N S , R. & H O F F M A N , T. () Wildlife damage
                                                                                                                          management research needs: perceptions of scientists, wildlife
             which may provide easily accessible food for bears, resulting
                                                                                                                          managers, and stakeholders of the USDA/Wildlife Services program.
             in an increased number of damage events (Northrup et al.,                                                    International Biodeterioration & Biodegradation, , –.
             ). Mitigation efforts should therefore be focused on                                                 B U R N H A M , K.P. & A N D E R S O N , D.R. () Model Selection and
             agricultural landscapes near protected areas and forests.                                                    Multimodel Inference: A Practical Information-theoretic Approach.
                The robustness of our analysis depends on the quality of                                                  Springer Verlag, New York, USA.
             the data on damage reports, and not all damage claims may                                                C A N , Ö.E., D’C R U Z E , N., G A R S H E L I S , D.L., B E E C H A M , J. &
                                                                                                                          M AC D O N A L D , D.W. () Resolving human–bear conflict: a global
             be reported to the national authorities. However, given that                                                 survey of countries, experts, and key factors. Conservation Letters, ,
             compensation is paid only if damage is reported, we are con-                                                 –.
             fident that the hotspot we identified is based on a real pat-                                            C AO , K.-A.L., R O S S OW , D., R O B E R T -G R A N I É , C. & B E S S E , P. () A
             tern of damage clustering in this area.                                                                      sparse PLS for variable selection when integrating omics data.
                Although our analysis has the potential to improve                                                        Statistical Applications in Genetics and Molecular Biology, , –.
                                                                                                                      C A R R A S C A L , L.M., G A LV Á N , I. & G O R D O , O. () Partial least
             human–wildlife coexistence in Croatia, our results also
                                                                                                                          squares regression as an alternative to current regression methods
             show that only a small fraction of the data variability is ex-                                               used in ecology. Oikos, , –.
             plained by our models. Other factors that were not considered                                            CLC () CORINE Land Cover . European Environment
             may influence patterns of damage, indicating the need for a                                                  Agency, Copenhagen, Denmark. Http://land.copernicus.eu/pan-
             broader analysis of the determinants of damage in Croatia.                                                   european/corine-land-cover/clc-/view [accessed  June ].
                                                                                                                      C O O R D I N AT I O N B O A R D F O R B E A R M A N A G E M E N T I N A U S T R I A ()
             The close relation between those making claims (i.e. farmer-
                                                                                                                          Bears in Austria – A Management Plan. Reviewed version .
             hunters) and the organizations responsible for compensation                                                  WWF Österreich, Wien, Austria.
             payments presents an opportunity to reduce conflict further                                              D E Č A K , Đ., F R KO V I Ć , A., G R U B E Š I Ć , M., H U B E R , Đ., I V I Č E K , B.,
             through prevention. Such an approach would help maintain                                                     K U L I Ć , B. et al. () Brown Bear Management Plan for the
             economic sustainability and cultural traditions in Croatia                                                   Republic of Croatia. Ministry of Agriculture, Forestry and Water
                                                                                                                          Management, Zagreb, Croatia.
             while also protecting the brown bear population.
                                                                                                                      D O B S O N , A.J. () An Introduction to Generalized Linear Models.
                                                                                                                          Chapman & Hall, Boca Raton, USA.
             Acknowledgements This study was supported by the LIFE                                                    E U R O S T AT () Agriculture, Forestry and Fishery Statistics – 
             DINALP BEAR project, EURONATUR and Bernd Thies                                                               Edition. Publications Office of the European Union, Luxembourg.
             Foundation, and the Ministry of Agriculture of Croatia. DH, LMR                                          F O U R L I , M. () Compensation for Damage Caused by Bears and
             and CF were supported financially by the University of Aveiro                                                Wolves in the European Union: Experiences from LIFE-Nature
             (Department of Biology), CESAM (UID/AMB/50017), and FCT/                                                     Projects. European Commission, Luxembourg City, Luxembourg.
             MEC through national funds, and co-funding by the FEDER within                                           G Ö T Z , O., L I E H R -G O B B E R S , K. & K R A F F T , M. () Evaluation of
             the PT2020 Partnership Agreement and Compete 2020.                                                           structural equation models using the partial least squares (PLS)
                                                                                                                          approach. In Handbook of Partial Least Squares, Springer
             Author contributions Data collection: DH and SR; data analysis:                                              Handbooks of Computational Statistics (eds V.E. Vinzi, W.W. Chin,
             DH, SR and LMR; writing and revisions: DH, LMR, CF, ĐH and                                                   J. Henseler & H. Wang), pp. –. Springer Verlag, Berlin and
             SMW; project coordination: ĐH.                                                                               Heidelberg, Germany.
                                                                                                                      G U N T H E R , K.A., H A R O L D S O N , M.A., F R E Y , K., C A I N , S.L., C O P E L A N D ,
             Conflicts of interest None.                                                                                  J. & S C H WA R T Z , C.C. () Grizzly bear–human conflicts in the
                                                                                                                          Greater Yellowstone ecosystem, –. Ursus, , –.
                                                                                                                      H A N S E N , A.J. () Contribution of source–sink theory to protected
             Ethical standards This research complies with the Oryx Code of
                                                                                                                          area science. In Sources, Sinks, and Sustainability Across Landscapes
             Conduct for authors.
                                                                                                                          (eds J. Liu, V. Hull, A. Morzillo & J. Wiens), pp. –. Cambridge
                                                                                                                          University Press, Cambridge, UK.
             References                                                                                               H U B E R , D. & R O T H , H.U. () Movements of European brown bears
                                                                                                                          in Croatia. Acta Theriologica, , –.
             B A N N E R , K.M. & H I G G S , M.D. () Considerations for assessing                                H U B E R , Đ., K U S A K , J., M A J I Ć -S K R B I N Š E K , A., M A J N A R I Ć , D. &
                model averaging of regression coefficients. Ecological Applications,                                      S I N D I Č I Ć , M. (a) A multidimensional approach to managing
                , –.                                                                                                the European brown bear in Croatia. Ursus, , –.
             B A R T O N , K. () MuMIn: Multi-Model Inference. R package version                                  H U B E R , Đ., J A K Š I Ć , Z., F R KO V I Ć , A., Š T A H A N , Ž., K U S A K , J.,
                ... Http://CRAN.R-project.org/package=MuMIn [accessed                                                M A J N A R I Ć , D. et al. (b) Brown Bear Management Plan for the
                June ].                                                                                               Republic of Croatia. Ministry of Regional Development, Forestry
             B A U T I S T A , C., N AV E S , J., R E V I L L A , E., F E R N Á N D E Z , N., A L B R E C H T , J.,       and Water Management, Directorate for Hunting and Ministry of
                S C H A R F , A.K. et al. () Patterns and correlates of claims for                                    Culture, Directorate for the Protection of Nature, Zagreb, Croatia.

                                                                                                           Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
Brown bear damage patterns in Croatia                                519

           J E R I N A , K., J O N O Z O V I Č , M., K R O F E L , M. & S K R B I N Š E K , T. ()                  Statistical Computing, Vienna, Austria. Http://www.R-project.org
               Range and local population densities of brown bear Ursus arctos in                                      [accessed  June ].
               Slovenia. European Journal of Wildlife Research, , –.                                       R I G G , R., F I N Ď O , S., W E C H S E L B E R G E R , M., G O R M A N , M.L.,
           K AC Z E N S K Y , P. () Large carnivore depredation on livestock in                                    S I L L E R O -Z U B I R I , C. & M AC D O N A L D , D.W. () Mitigating
               Europe. Ursus, , –.                                                                               carnivore–livestock conflict in Europe: lessons from Slovakia. Oryx,
           K A R A M A N L I D I S , A., S A N O P O U LO S , A., G E O R G I A D I S , L. & Z E D R O S S E R ,       , –.
               A. () Structural and economic aspects of human–bear conflicts                                   R O Y , P.P. & R O Y , K. () On some aspects of variable selection for
               in Greece. Ursus, , –.                                                                          partial least squares regression models. QSAR & Combinatorial
           K AV Č I Č , I., A D A M I Č , M., K AC Z E N S K Y , P., K R O F E L , M. & J E R I N A , K.               Science, , –.
               () Supplemental feeding with carrion is not reducing brown                                      S A N C H E Z , G. () plsdepot: Partial Least Squares (PLS) data analysis
               bear depredations on sheep in Slovenia. Ursus, , –.                                             methods. R package version ... Http://CRAN.R-project.org/
           K N O T T , E.J., B U N N E F E L D , N., H U B E R , Đ., R E L J I Ć , S., K E R E Ž I , V. &              package=plsdepot [accessed  June ].
               M I L N E R -G U L L A N D , E.J. () The potential impacts of changes in                        S C H W E R D T N E R , K. & G R U B E R , B. () A conceptual framework
               bear hunting policy for hunting organisations in Croatia. European                                      for damage compensation schemes. Biological Conservation, ,
               Journal of Wildlife Research, , –.                                                                –.
           K O C I J A N , I. & H U B E R , Đ. () Conservation Genetics of Brown Bears                         S E R V H E E N , C., H E R R E R O , S. & P E Y T O N , B. (eds) () Bears: Status
               in Croatia. Final Report. Project: Gaining and Maintaining Public                                       Survey and Conservation Action Plan. International Union for
               Acceptance of Brown bear in Croatia (BBI-Matra// through                                         the Conservation of Nature and Natural Resources, Gland,
               ALERTIS), Zagreb, Croatia.                                                                              Switzerland.
           K O M S T A , L. () Outliers: Tests for outliers. R package version ..                           S K R B I N Š E K , T., B R A G A L A N T I , N., C A L D E R O L L A , S., G R O F F , C., H U B E R ,
               Http://CRAN.R-project.org/package=outliers [accessed  June ].                                      D., K AC Z E N S K Y , P. et al. ()  Annual Population Status
           K R U U K , H. () Hunter and Hunted. Relationships between Carnivores                                   Report for Brown Bears in Northern Dinaric Mountains and Eastern
               and People. Cambridge University Press, Cambridge, UK.                                                  Alps. Life DinAlp bear report (LIFE NAT/SI/), Zagreb,
           M A H M O U D , M., Q A S H Q A E I , A.T., M A R A S H I , M. & N E J AT , F. ()                       Croatia.
               Seasonal human–brown bear conflicts in northern Iran:                                               S O N G C H I T R U K S A , P. & Z E N G , X. () Getis–Ord spatial statistics to
               implications for conservation. Zoology and Ecology, , –.                                        identify hot spots by using incident management data.
           M A J I Ć , A., M A R I N O T A U S S I G D E B O D O N I A , A., H U B E R , Đ. &                          Transportation Research Record: Journal of the Transportation
               B U N N E F E L D , N. () Dynamics of public attitudes toward bears                                 Research Board, , –.
               and the role of bear hunting in Croatia. Biological Conservation, ,                              S V E Č N J A K , L., H E G I Ć , G., K E Z I Ć , J., T U R Š I Ć , M., D R A Ž I Ć , M.M.,
               –.                                                                                              B U B A LO , D. & K E Z I Ć , N. () The state of beekeeping in Croatia.
           N O R T H R U P , J.M., S T E N H O U S E , G.B. & B O Y C E , M.S. () Agricultural                     Central European Journal of Agriculture, , –.
               lands as ecological traps for grizzly bears. Animal Conservation, ,                               S W E N S O N , J.E., G E R S T L , N., D A H L E , B. & Z E D R O S S E R , A. ()
               –.                                                                                                Action plan for the conservation of the brown bear (Ursus arctos) in
           N Y H U S , P., F I S C H E R , H., M A D D E N , F. & O S O F S K Y , S. () Taking the                 Europe. Nature and Environment Series (Council of Europe,
               bite out of wildlife damage. Conservation in Practice, , –.                                        Strasbourg), , –.
           N Y H U S , P., O S O F S K Y , S., F E R R A R O , P., F I S C H E R , H. & M A D D E N , F.           T R E V E S , A. & K A R A N T H , U.K. () Human–carnivore conflict and
               () Bearing the costs of human–wildlife conflict: the challenges                                     perspectives on carnivore management worldwide. Conservation
               of compensation schemes. In People and Wildlife: Conflict or                                            Biology, , –.
               Coexistence? (eds R. Woodroffe, S. Thirgood & A. Rabinowitz), pp.                                   T R O U W B O R S T , A. () Managing the carnivore comeback:
               –. Cambridge University Press, Cambridge, UK.                                                     international and EU species protection law and the return of lynx,
           O R D , J.K. & G E T I S , A. () Local spatial autocorrelation statistics:                              wolf and bear to Western Europe. Journal of Environmental Law, ,
               distributional issues and an application. Geographical Analysis, ,                                    –.
               –.                                                                                            W I L S O N , S.M., G R A H A M , J.A., M AT T S O N , D.J. & M A D E L , M.J. ()
           P E T R A M , W., K N A U E R , F. & K A C Z E N S K Y , P. () Human influence                          landscape conditions predisposing grizzly bears to conflict on
               on the choice of winter dens by European brown bears in Slovenia.                                       private agricultural lands in the western USA. Biological
               Biological Conservation, , –.                                                                  Conservation, , –.
           P O S I L L I C O , M., M E R I G G I , A., P A G N I N , E., L O VA R I , S. & R U S S O , L.          Z A R , J.H. () Biostatistical Analysis. Pearson Prentice Hall, Upper
               () A habitat model for brown bear conservation and land                                             Saddle River, London, UK.
               use planning in the central Apennines. Biological Conservation, ,                                Z E D R O S S E R , A., D A H L E , B., S W E N S O N , J.E. & G E R S T L , N. () Status
               –.                                                                                                and management of the brown bear in Europe. Ursus, , –.
           R D E V E LO P M E N T C O R E T E A M () R: A Language and                                         Z U U R , A.F., I E N O , E.N. & S M I T H , G.M. () Analysing Ecological
               Environment for Statistical Computing. R Foundation for                                                 Data. Springer, New York, USA.

           Oryx, 2020, 54(4), 511–519 © 2018 Fauna & Flora International doi:10.1017/S0030605318000236
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 15 Feb 2021 at 13:38:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318000236
You can also read