Environmental predictors of livestock predation: a lion's tale
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Environmental predictors of livestock predation:
a lion's tale
J . A . D . R O B E R T S O N , M . R O O D B O L , M . D . B O W L E S , S . G . D U R E S and J . M . R O W C L I F F E
Abstract Negative interactions between people and large Introduction
carnivores are common and will probably increase as the
human population and livestock production continue to
expand. Livestock predation by wild carnivores can signifi- H uman–wildlife interactions vary in frequency, inten-
sity, and on a continuum from positive to negative
(Soulsbury & White, ). Negative interactions can pose
cantly affect the livelihoods of farmers, resulting in re-
taliatory killings and subsequent conflicts between local significant threats to human welfare and livelihoods, and
communities and conservationists. A better understanding to ecosystem structure and function (Estes et al., ;
of livestock predation patterns could help guide measures Kansky et al., ). They are often complex, encompassing
to improve both human relationships and coexistence with an array of species and situations, each with unique factors
carnivores. Environmental variables can influence the in- and thus potential solutions (Dickman, ). Conflicts
tensity of livestock predation, are relatively easy to monitor, resulting from negative human–wildlife interactions are
and could potentially provide a useful predictive framework likely to be exacerbated in the future with further growth
for targeting mitigation. We chose lion predation of live- of the human population and ongoing destruction of
stock as a model to test whether variations in environmental habitats for wildlife.
conditions trigger changes in predation. Analysing years Large carnivores can cause substantial direct and indirect
of incident reports for Pandamatenga village in Botswana, costs to humans and are particularly prevalent in adverse
an area of high human–lion conflict, we used generalized human–wildlife interactions because they often predate live-
linear models to show that significantly more attacks coin- stock and have large home ranges that are likely to overlap
cided with lower moonlight levels and temperatures, and at- with human-dominated areas (Macdonald & Sillero-Zubiri,
tack severity increased significantly with extreme minimum ; Inskip & Zimmermann, ). Conserving large
temperatures. Furthermore, we found a delayed effect of carnivores is vital because they are keystone species that
rainfall: lower rainfall was followed by a significantly in- act as umbrella and flagship species for wider biodiver-
creased severity of attacks in the following month. Our sity protection (Loveridge et al., ; Macdonald &
results suggest that preventative measures, such as introdu- Loveridge, ), and many are deeply intertwined with
cing deterrents or changing livestock management, could be human cultures (Kellert et al., ). However, most (%)
implemented adaptively based on environmental condi- of the largest carnivores now have decreasing populations
tions. This could be a starting point for investigating similar (Ripple et al., ). For example, although negative
effects in other large carnivores, to reduce livestock attacks interactions between humans and African lions Panthera
and work towards wider human–wildlife coexistence. leo have happened for millennia (Loveridge et al., ),
the lion population (%) and range (%) have declined
Keywords Carnivores, human–wildlife conflict, lion, live- dramatically in recent decades (Bauer et al., ). Habitat
stock predation, moon, Panthera leo, rainfall, temperature destruction and land-use changes, resulting from the cessa-
Supplementary material for this article is available at tion of traditional sustainable land-use practices following
https://doi.org/./S European colonization of African countries, are primarily
responsible for these declines (Clover & Eriksen, ;
Bauer et al., ). As a consequence of their now restricted
range and diminished population, the future survival of the
African lion may depend upon coexistence with farmers in
human-dominated landscapes, because the effectiveness of
protected areas can be compromised when lions are drawn
out of these areas to replace those killed on the borders
(Loveridge et al., ; Macdonald & Loveridge, ).
J. A. D. ROBERTSON (Corresponding author) Silwood Park Campus, Imperial
College London, London, UK. E-mail joshrobertsoniwb@gmail.com Negative human–carnivore interactions will probably
M. ROODBOL and M. BOWLES Walking for Lions, Pandamatenga, Botswana
escalate in the future as global livestock production is pro-
jected to rise from Mt in – to Mt in ,
S. G. DURES and J. M. ROWCLIFFE Institute of Zoology, Zoological Society of
London, London, UK with much of this growth occurring in developing countries
Received April . Revision requested August . (Alexandratos & Bruinsma, ). This is a complex issue
Accepted September . First published online June . (Dickman, ; Madden & McQuinn, ), with both
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217
Environmental predictors of livestock predation 649
tangible and intangible costs to human communities and We investigated how temperature, moon phase and rain-
carnivores (Redpath et al., a; Dickman & Hazzah, fall affected the incidence and severity of lion predation
; Kansky et al., ). In particular, the economic of livestock over a -year period in an area with frequent
costs of predation for livestock farmers are significant negative human–lion interactions. We hypothesized that
(Butler, ; Patterson et al., ) and can lead to () livestock predation would increase with decreasing
substantial losses of annual household income (Wang & temperature, () livestock predation would increase with
Macdonald, ). Consequently, retaliatory killings are decreasing moonlight levels, and () rainfall would have
common and lead to conflicts between conservationists no significant effect on livestock predation, given that the
and livestock farmers (Redpath et al., b). study area is surrounded by protected areas.
Examining the factors that affect livestock predation can
help us understand patterns and subsequently guide mea- Study area
sures to reduce attacks, foster relationships between con-
servationists and farmers, and work towards coexistence The study area (Fig. ) covers c. , km around
of farmers and predators (Carvalho et al., ). Livestock Pandamatenga village in north-east Botswana, in the
predation risk is influenced by multiple variables, including Chobe district near the border with Zimbabwe and close
livestock abundance, farming techniques, mitigation mea- to Hwange National Park. Pandamatenga’s human popula-
sures, presence of trophy hunting, changes in laws and tion increased from , in (African Development
their implementation, proximity to human settlements Bank, ) to c. , in (Central Statistics Office of
and protected areas (Yu et al., ; Van Bommel et al., Botswana, ). It is one of Botswana’s least arid areas,
), natural prey density and environmental conditions with a mean annual rainfall of mm, almost all of
(Sunquist & Sunquist, ; Polisar et al., ; Azevedo & which occurs during October–April, with a peak during
Murray, ; Tortato et al., ). Environmental condi- December–February (African Development Bank, ).
tions such as temperature, light level and rainfall are rela- Pandamatenga is situated between several protected
tively easy to monitor and have overarching impacts on areas that support c. lion prides comprising at least in-
livestock predation by carnivores. It is plausible that varia- dividuals. Farms are small subsistence operations and all
tions in such factors could trigger changes in predation pat- livestock are kept in protective enclosures (kraals) at night
terns, and could thus be used as predictors of the likelihood and released for grazing during the day; the robustness of
of attacks. We chose livestock predation by lions as a model the kraals, presence of any other mitigation methods and
interaction to test this idea. grazing regimes differ between farmers.
Environmental conditions, particularly temperature and
rainfall, have significant effects on lion biology and demo- Methods
graphy (Celesia et al., ). Rainfall has been associated
with both increases (Patterson et al., ; Kuiper et al., Incident reports
) and decreases in lion attacks on livestock (Butler,
; Schiess-Meier et al., ), which are typically related Botswana provides a state-funded compensation scheme in
to prey abundance. Compared to non-protected landscapes, which farmers file a report when they lose livestock to lions
protected areas usually have higher prey levels that remain (Hemson et al., ). The Department of Wildlife and
relatively constant over time in non-migratory systems National Parks provided us with reports on incidents that
(Macdonald & Loveridge, ). This suggests rainfall is took place during January –April . We believe
less likely to influence rates of livestock predation in regions these reports provide comprehensive coverage of lion–live-
surrounded by protected areas. Although temperature is stock incidents during this period because farmers must file
thought to have no influence on lion predation of livestock a report to claim compensation for attacks, and farmers in
(Patterson et al., ), lions are more active during colder Pandamatenga lack the financial capacity to leave these
periods and more nocturnal in areas with higher mean an- unreported. The data included reports on incidents,
nual temperatures (Hayward & Slotow, ); temperature each with information on date, location, source, livestock
also affects food intake by lions (West & Packer, ). owner, the species and number of livestock involved, and
Because hunting for prey increases body temperature in whether the incident took place inside or outside a kraal.
lions, and their biology leaves them vulnerable to over-
heating, we expected livestock predation levels to reflect Data preparation
their preference for lower external temperatures. Lions
prefer periods without moonlight for hunting wild prey Three response variables were explored: incident occurrence
(Schaller, ), and travel closer to livestock at lower moon- (a measure of attack likelihood), the number of livestock
light levels (Oriol-Cotterill et al., ), probably because of involved in incidents (a measure of attack severity), and
the reduced risk of being seen by people and prey. the total cost of incidents (an alternative measure of attack
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217
650 J. A. D. Robertson et al.
We accessed monthly rainfall data (mm) from the me-
teorological office in Pandamatenga for July –March
. Daily moon phase data during January –
April were downloaded from USNO ().
According to results from Packer et al. (), three moon
phases were defined: the full moon and subsequent days
(phase ), the days prior to the full moon (phase ),
and the intermediate days when the moon is least visible
(phase ). Moonlight levels are lowest during phase and
highest during phases and ; however, during phase the
moon does not rise until after sunset and thus leaves a per-
iod of darkness, whereas during phase the moon is above
the horizon at sunset and there is no such interlude. We
obtained data on temperature for Pandamatenga during
January – April from NOAA (). This in-
cluded daily values for minimum and maximum tempera-
ture, and monthly values for extreme daily maximum and
minimum, monthly mean maximum and minimum, and
monthly mean temperatures. We used linear interpolation
FIG. 1 The study site around Pandamatenga village in northeast to fill in minimum and maximum temperature va-
Botswana, showing the area from which the analysed incident lues that were missing from the daily data (of a total of ,
reports on livestock predation by lions Panthera leo originated. values), with average data gaps being . and . days,
respectively; linear interpolation was deemed to be more
severity, reflecting farmers’ likely perception of the events). accurate than polynomial, based on plots of the data.
All three response variables were expressed on a daily, month- Monthly temperature and rainfall data were shifted to
ly, and annual scale. We used a daily scale to investigate the match the response variables and months ahead, to
effect of temperature and moon phase on response variables. test for any delayed response; extreme daily temperatures
Daily data covered each day during January – April were not used in these models because it is unlikely that
, and included incidents, non-incidents, and the asso- an extreme temperature on a single day would affect the
ciated response and environmental variables; we removed incidents in the following months.
values of zero for the analysis of livestock attacked per inci-
dent and associated cost. We used a monthly scale to investi- Data analysis
gate the immediate and delayed effect of temperature and
rainfall on monthly totals of response variables during We used R .X for all data analysis (R Development Core
January –April . We used an annual scale to examine Team, ). To test our hypotheses and investigate the
large-scale patterns in response variables during –; effect of environmental variables on response variables, we
data from were removed from the annual analysis used generalized linear models and assumed the follow-
because they were only available until April that year. ing error distributions: binomial for the occurrence of an
Monetary cost in Botswana Pula (BWP) was assigned to incident, Poisson for the number of incidents, livestock
each incident based on current compensation scheme va- attacked per month and livestock attacked per incident,
lues; as of these were set in BWP as for a goat, and Gaussian for cost (log) per incident and month. We used
, for a calf, , for a pig, , for a heifer, cow, bul- generalized linear models for daily data to assess the effect
lock, or oxen, and , for a bull. Costs were assigned to the of temperature and moon phase on response variables, on
incidents irrespective of whether an animal was killed or in- monthly data to assess the effect of temperature and rainfall,
jured, because information on the severity of the injuries or and on shifted monthly data to assess delayed effects of
any livestock fatalities was not included in the reports. Five temperature and rainfall. To control for interdependence
reported incidents concerned lions being translocated from between variables, we excluded variables correlated by
farmers’ lands; these were removed because they did not in- r . .. When our models assuming Poisson error distri-
volve attacks on livestock. Incidents involving dogs (n = ) butions showed a residual deviance higher than residual de-
and chickens (n = ) were included as incidents, but not in grees of freedom, we compensated for this overdispersion by
the number of livestock attacked or the cost of incidents as assuming quasi-Poisson errors. We followed a step-by-step
neither are covered under the compensation scheme: dogs procedure to find the model with the best fit based on
are not livestock, and the single incident involving chick- Akaike’s information criterion (AIC) and the residual
ens could have exerted undue leverage on the results. deviance relative to degrees of freedom.
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217
Environmental predictors of livestock predation 651
TABLE 1 Summary of the best fit generalized linear models examining the effect of moon phase and temperature on incidents of lion
Panthera leo predation on livestock and their severity in Pandamatenga. Significant effects are indicated with *.
Response variable (error distribution) Environmental variable1 Coefficient estimate SE P Residual deviance (df)
Incident (Binomial) Intercept −1.195 0.190 , 0.001* 2,099.2 (2353)
Moon phase 2 −0.278 0.140 0.050*
Moon phase 3 0.147 0.130 0.262
TMin −0.026 0.011 0.017*
Livestock attacked (Poisson) Intercept 0.713 0.140 , 0.001* 280.0 (374)
Moon phase 2 −0.040 0.110 0.711
Moon phase 3 −0.032 0.096 0.740
TMin −0.017 0.008 0.032*
Cost (log) (Gaussian) Intercept 7.923 0.160 , 0.001* 300.7 (375)
Moon phase 2 −0.067 0.120 0.575
Moon phase 3 −0.076 0.100 0.476
TMin −0.008 0.009 0.382
TMin, minimum temperature (°C) on a given day.
Results
Over the years of the study there were lion attacks
involving livestock at a total cost of BWP ,,
(GBP ,; Supplementary Table ).
Temperature and moon phase
We could not use minimum and maximum temperature in
the same generalized linear model because of intercorrel-
ation, and minimum temperature produced the best-fit
model for each response variable. Our best-fit generalized
linear models (Table ) showed that the incidence of attacks
FIG. 2 The predicted probability of a lion attack on livestock in on livestock and the number of livestock attacked per
Pandamatenga based on the minimum temperature (°C) of a given
event both increased with decreasing minimum temperature
day under different moon phases: phase is the full moon and
subsequent days, phase is the days prior to the full moon, (Fig. ). We also found this pattern for maximum tempera-
and phase includes the remaining days around the new moon. ture, although the effect was weaker (Supplementary Fig. ).
Based on the coefficient for minimum temperature, holding
Because we split moon phase into three categories, we moon phases at a fixed value, there was a .% increase in
used a general linear hypothesis test to assess differences the odds of an incident with every °C decrease. Incidents
amongst them. We calculated the predicted daily probabil- were more likely to occur in moon phases (around new
ity of an attack on the back-transformed response scale, moon) and (post-full moon) compared to moon phase
for a sequence of temperatures at each moon phase, (pre-full moon); however, there were no differences between
and plotted responses with % confidence intervals; we moon phases for either the number of livestock attacked per
extended the minimum temperature range to −– °C incident or associated cost. Odds ratios showed that the
(from −.–. °C), and maximum temperature range likelihood of an incident occurring during moon phases
to – °C (from .–. °C) to investigate how and was . and . times higher, respectively, than
potential future changes in extreme temperatures might during phase (Supplementary Table ). No environmental
affect livestock predation. variables were significant predictors of cost. These results
We used a Kruskal–Wallis H test to assess differences in support our hypotheses and that livestock predation
the number of livestock involved and cost of incidents be- would increase with decreasing temperature and moonlight.
tween the years based on a per incident basis and on month-
ly summaries. Seasons were defined as dry (May–October)
and wet (November–April). We compared monthly sums of Temperature and rainfall
attacks inside and outside kraals and between seasons for
all response variables using a Welch two-sample t test and All monthly temperature variables were correlated and we
visualized them with box plots. selected only the temperature variable that produced the
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217652 J. A. D. Robertson et al.
TABLE 2 Summary of the best-fit generalized linear models examining the effect of monthly rainfall and temperature on lion–livestock
incidents and their severity in Pandamatenga. Significant effects are indicated with *.
Response variable (error distribution) Environmental variable1 Coefficient estimate SE P Residual deviance (df)
Incidents (Quasi-Poisson) Intercept 2.456 0.470 , 0.001* 84.2 (30)
MMeanMin −0.070 0.035 0.058
Rainfall 0.003 0.003 0.239
Livestock attacked (Quasi-Poisson) Intercept 2.894 0.270 , 0.001* 122.2 (30)
ExMinTemp −0.103 0.033 0.004*
Rainfall 0.004 0.003 0.205
Cost (log) (Gaussian) Intercept 9.883 2.020 , 0.001* 244.6 (30)
MMeanMin −0.039 0.140 0.783
Rainfall −0.015 0.001 0.143
MMeanMin, monthly mean minimum temperature (°C); ExMinTemp, extreme minimum daily temperature (°C).
temperature months earlier (Supplementary Fig. ). Con-
trary to our hypothesis that rainfall would not influence
attacks in the context of nearby protected areas, we found
high levels of rainfall decreased the number of livestock
attacked and the associated cost in the following month
(Fig. ). Rainfall had no significant -month lag effect on
response variables.
Temporal patterns and the effectiveness of kraals
Attacks on livestock tended to be more frequent and severe in
the dry season (Supplementary Fig. ). However, a Kruskal–
Wallis H test showed there was no significant difference
between months in the number of incidents (χ() = .,
FIG. 3 The effect of extreme minimum daily temperature (°C) on P = .), livestock attacked (χ() = ., P = .) or
the number of livestock attacked by lions in Pandamatenga per cost (χ() = ., P = .). Similarly, a Kruskal–Wallis
month, showing the fitted generalized linear model and Wald H test showed there was no significant difference
% confidence intervals based on standard errors. between years in the number of incidents (χ() = .,
P = .), livestock attacked (χ() = ., P = .), or
best fit for each model (Table ). The monthly number of cost (χ() = ., P = .). A Welch two-sample t test
livestock attacked increased significantly with decreasing showed that significantly more incidents (t = −.,
extreme minimum temperature (Fig. ), and marginally df = ., P , .), involving more livestock
with decreasing extreme maximum daily temperature (t = −., df = ., P , .), at a higher cost (t = −.,
(Supplementary Fig. ); this result again supports our df = ., P , .), occurred outside kraals, based on
hypothesis that decreasing temperature would increase monthly summaries (Supplementary Fig. ).
livestock predation. Although decreasing monthly mean
minimum temperatures increased the number of incidents
Discussion
and associated costs, these effects were not significant.
Rainfall was not predictive of any response variable, which Although temperature and moon phase are known to affect
supports our hypothesis that rainfall has no effect on live- lion activity (Hayward & Slotow, ; Celesia et al., ;
stock predation in locations surrounded by protected areas. Packer et al., ; Oriol-Cotterill et al., ), our study is
the first to show these variables have a significant effect
on the occurrence and severity of lion attacks on livestock.
Delayed effects of temperature and rainfall
Temperature had no significant -month delayed effect on Effects of temperature
response variables, nor -month delayed effect on incidents
or cost (Table ). However, the number of livestock attacked Our hypothesis that livestock predation would increase with
was negatively associated with mean monthly maximum decreasing temperature was supported by all models. There
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217Environmental predictors of livestock predation 653
TABLE 3 Summary of the best-fit generalized linear models examining the delayed effect of temperature and rainfall on lion–livestock
incidents and their severity in Pandamatenga. Significant effects are indicated with *.
Response variable Environmental Coefficient Residual
(error distribution) Delay (months) variable1 estimate SE P deviance (df)
Incidents (Quasi-Poisson) 1 Intercept 3.595 1.410 0.017* 82.73 (30)
MeanMaxTemp −0.064 0.047 0.179
Rainfall −0.001 0.002 0.597
2 Intercept 1.103 0.550 0.053 87.70 (29)
MeanMinTemp 0.034 0.370 0.366
Rainfall −0.002 0.003 0.437
Livestock attacked (Quasi-Poisson) 1 Intercept 4.688 1.260 , 0.001* 113.48 (30)
MeanMaxTemp −0.081 0.042 0.062*
Rainfall −0.005 0.002 0.047*
2 Intercept 5.585 1.450 , 0.001* 128.24 (29)
MeanMaxTemp −0.117 0.049 0.023*
Rainfall 0.0003 0.002 0.886
Cost (log) (Gaussian) 1 Intercept 14.254 5.190 0.010* 225.06 (30)
MeanMaxTemp −0.152 0.167 0.368
Rainfall −0.018 0.007 0.024*
2 Intercept 19.942 5.580 0.001* 245.50 (29)
MeanMaxTemp −0.365 0.179 0.051
Rainfall 0.001 0.008 0.870
MeanMaxTemp, mean monthly maximum temperature (°C); MeanMinTemp, mean monthly minimum temperature (°C).
have been few previous studies on the effect of temperature and the insulating properties of their manes, and males feed
on livestock predation by carnivores. Patterson et al. () significantly less in warmer months, whereas food intake of
found no correlation between temperature and the fre- females remains unchanged at higher temperatures (West &
quency or severity of lion attacks on livestock. However, Packer, ). Most lions predating livestock are male,
their analysis only included temperature data for most days and thus males are probably overrepresented in our data
during year whereas we were able to analyse years of data. (Macdonald & Loveridge, ). Although hunting livestock
Livestock predation is affected by husbandry practices, may require less physical exertion than hunting wild
which vary between individuals and on a spatio-temporal prey, this could conceivably be outweighed by the fear of
scale (Kgathi et al., ). For example, some farmers may human activities in areas with frequent negative human–
leave their cattle out for grazing at night during warmer per- lion interactions. For example, the body temperature of
iods, and others may put less effort into preventing livestock cheetahs significantly increases after a successful hunt
attacks during colder periods to avoid uncomfortable condi- because of the stress induced from remaining vigilant for
tions, resulting in more incidents. However, we have no data a dominant predator, rather than the physical exertion of
on individual farmers’ husbandry strategies to assess this. the hunt itself (Hetem et al., ). Lions hunting livestock
We can, however, infer from our results that lion physi- around pastoral lands exhibit a wariness of human activity,
ology and behaviour influence livestock predation. Lions altering their behaviour and moving faster than usual
must maintain their core body temperature at – °C, (Valeix et al., ; Oriol-Cotterill et al., ). It is therefore
but have few behavioural mechanisms to facilitate this at conceivable that lions will be in an alarmed state after a suc-
high ambient temperatures (Willmer et al., ). Given cessful livestock hunt, resulting in higher body temperatures
that a successful livestock hunt increases a lion’s body and stress hyperthermia (Meyer et al., ; Hetem et al.,
temperature (through locomotion, stress and feeding) and ), and thus have a preference for predating livestock
a lion’s biology leaves it vulnerable to overheating, it is during colder periods. Further increasing livestock preda-
probable that lions favour lower and avoid higher ambient tion during extremely low temperatures may be a behav-
temperatures for livestock predation. Overheating when ioural strategy to avoid critically low body temperatures
hunting is a significant danger for lions because they have (Stryker, ), as lions require more energy for thermo-
large amounts of heat-producing muscle tissue and a low regulation during these periods, and large meat-based
surface area to volume ratio, and because locomotion can meals increase body temperature significantly (West &
increase heat production to – times that of the basal Packer, ; Pough et al., ).
metabolic rate (Pough et al., ). Overheating is probably Rather than temperature having a direct effect on preda-
a greater problem for males because of their larger body size tion, it could be that variations in wild prey abundance
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217654 J. A. D. Robertson et al.
Nxai Pan National Park in the wet season may decrease
prey availability around Pandamatenga during that time.
However, our results show incident number and severity
to be generally higher in the dry season, so it is unlikely
this migration affects attacks around Pandamatenga.
Furthermore, high availability of wild prey has been
shown to both increase (Stahl et al., ; Treves et al.,
) and decrease (Hemson, ; Valeix et al., )
levels of livestock predation; increases in livestock
predation may occur because higher wild prey levels
can sustain greater predator populations (Macdonald &
Loveridge, ).
Effects of moon phase
Our findings agreed with our hypothesis that livestock pre-
dation would increase with decreasing moonlight levels.
This is consistent with research indicating that lions travel
closer to kraals at lower moonlight levels (Oriol-Cotterill
et al., ), and that attack rates of lions on people was
– times lower in the days before the full moon
compared to the days after, when the moon rises after
sunset, resulting in a period of darkness (Packer et al.,
). Lions move more slowly and their movement paths
are more tortuous when moonlight levels are high, indicat-
ing more caution (Oriol-Cotterill et al., ), and we infer
that lions utilize periods of lower moonlight for hunting
livestock because it is easier to avoid detection by people
and potential prey.
Effects of rainfall
Rainfall had no immediate effect on the number of inci-
dents or their severity, which supports our hypothesis that
livestock predation should be unrelated to rainfall in the
context of stable wild prey numbers within nearby protected
areas. However, attacks were significantly less severe in the
month following increased rainfall, which contradicts this
hypothesis. Patterson et al. (), in a study carried out
in Tsavo National Park, found no delayed relationship
between rainfall and predation with lags of – months
FIG. 4 The effect of the sum of monthly rainfall on (a) the but found attacks to be most frequent in the wet season.
number of predation incidents, (b) the number of livestock
Pandamatenga’s different natural prey densities and land-
attacked, and (c) the log associated cost of attacks in
Pandamatenga in the following month, showing the fitted scape could be responsible for this discrepancy between
generalized linear model lines and Wald % confidence our results and the ones reported from Tsavo National
intervals based on standard errors. Park. Farmers take their livestock further from kraals
when vegetation density and water levels are low (Oriol-
Cotterill et al., ), leaving a higher number of livestock
coincide with seasonal temperature changes, as lion– more vulnerable to predation. Decreasing levels of rainfall
livestock predation is affected by the relative abundance of will probably lead to lower vegetation density and water
wild prey compared to livestock (Hemson, ). The mass availability in the following month, which could explain
movement of zebras (Naidoo et al., ), and potentially the delayed relationship between rainfall and attack
other ungulates (Harris et al., ), from Chobe river to severity.
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217Environmental predictors of livestock predation 655
Temporal patterns and the effectiveness of kraals typically coincide with the dry season in southern Africa.
Furthermore, although the effect of rainfall on livestock
The fact that there was no significant seasonal trend in predation varies in different areas, measures to prevent
the number or severity of attacks suggests that there was predator attacks should be increased in months following
relatively high prey availability throughout the year at lower levels of rainfall in locations surrounded by protect-
Pandamatenga, possibly because regions surrounded by ed areas. Such measures could include improvements in
protected areas typically have higher ungulate populations livestock management, such as keeping cattle within kraals
(Macdonald & Loveridge, ). We found significantly whenever possible and being more vigilant with herding and
more incidents occurred outside kraals, which is consistent guarding, using various deterrents (e.g. acoustic, visual or
with other research (Valeix et al., ) and suggests kraals chemical), using specialized guard dogs, or local dogs that
are effective to some extent in reducing attacks. act as a warning system, and employing human guardians
for livestock (Dickman, ). Each area will have different
potential for successful mitigations and carnivore con-
Limitations, and directions for future research
servation based on its ecological, sociological, economic
Although we have shown that environmental factors could and political profile, and the associated human population
have significant effects on livestock predation, this only pressure (Hemson, ; Dickman et al., ). Decreasing
facilitates an understanding of one part of a complex system predator attacks on livestock could help reduce the perse-
in which other variables, particularly human factors, are cution of large carnivores and maintain the essential
equally important (Oriol-Cotterill et al., ). Future work ecosystem services they provide, whilst improving relation-
should encompass a more holistic analysis of the variables ships between conservationists and farmers, and facilitating
affecting predation, including environmental, human, and human–carnivore coexistence.
spatial factors, to compare the effects of these broad
Acknowledgements JADR thanks Jocelyne Sze Shimin, George
categories on livestock predation by multiple large carni-
Powell and James Foley for their advice on statistical analyses;
vores in different locations. In addition, it would be of Joseph Shepherdson, Jim Humphries, James Adams, Jess Williams
value to investigate the effect of temperature and moon and Jacky Morrison for reviewing early drafts; and Hil and Jamie
phase specifically on livestock predation by other large Robertson for their support. All authors thank the Department of
carnivores in different ecosystems, to establish whether Wildlife and National Parks for providing the incident reports that
made this project possible.
these variables can act as indicators for guiding mitigation
more broadly. Author contributions Project conception and study design: JADR,
We had no information regarding the time of day of JMR, MR, SGD; writing: JADR; guidance on writing: JMR, SGD;
attacks, so although most attacks probably occur at night information on local governance, farming practices and lion
(Macdonald & Loveridge, ), inferences of high tempera- behaviour: MR, MDB, SGD; initial data collation: MDB; data
ture avoidance must be taken as general patterns exhibited analysis: JADR, JMR; figures and tables: JADR.
by the data rather than actual daily preferences shown by Conflicts of interest None.
lions. Confidence in our results would be strengthened
with data on natural prey density and lion predation levels Ethical standards All authors have abided by the Oryx Code of
on natural prey, which will have an impact on livestock pre- Conduct for contributors. This project did not involve ethical issues
dation. Additionally, moon brightness varies with weather with either human subjects or animals.
conditions, and our analysis did not consider cloud cover-
age, which is probably higher in the wet season. Future
References
work should incorporate these data and involve an analysis
of human activities and known attacks on both livestock and A F R I C A N D E V E LO P M E N T B A N K () Pandamatenga Agricultural
wild prey by focusing on activity patterns of individual lions. Infrastructure Development Project Appraisal Report. African
Development Bank, Tunis Belvedere, Tunisia.
A L E X A N D R A T O S , N. & B R U I N S M A , J. () World Agriculture
Management implications Towards /. ESA Working Paper No. -. Food and
Agriculture Organization of the United Nations, Rome, Italy.
Environmental variables are easy to monitor and can A Z E V E D O , F. & M U R R AY , D.L. () Evaluation of potential factors
provide practical guidelines for enhancing mitigation, to predisposing livestock to predation by jaguars. Journal of Wildlife
help reduce attacks from large carnivores and improve Management, , –.
farmer–carnivore coexistence. Considering the demonstra- B A U E R , H., P AC K E R , C., F U N S T O N , P., H E N S C H E L , P. & N OW E L L , K.
() Panthera leo. In The IUCN Red List of Threatened Species .
ted effect of all environmental variables we investigated,
Http://dx.doi.org/./IUCN.UK.-.RLTS.TA.en.
we recommend that livestock farmers increase protective B U T L E R , J.R.A. () The economic costs of wildlife predation on
measures for their livestock during colder periods, and livestock in Gokwe communal land, Zimbabwe. African Journal of
particularly when moonlight levels are low. Colder periods Ecology, , –.
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217656 J. A. D. Robertson et al.
C A R VA L H O , E.A.R., Z A R C O -G O N Z Á L E Z , M.M., M O N R OY -V I L C H I S , O. L O V E R I D G E , A.J., S E A R L E , A.W., M U R I N D A G O M O , F. & M A C D O N A L D ,
& M O R A T O , R.G. () Modeling the risk of livestock depredation D.W. () The impact of sport-hunting on the population
by jaguar along the Transamazon highway, Brazil. Basic and Applied dynamics of an African lion population in a protected area.
Ecology, , –. Biological Conservation, , –.
C E L E S I A , G.G., P E T E R S O N , A.T., J U L I A N , C. & G N O S K E , T.P. () L O V E R I D G E , A.J., P AC K E R , C. & D U T T O N , A. () Science and the
Climate and landscape correlates of African lion (Panthera leo) recreational hunting of lions. In Recreational Hunting, Conservation
demography. African Journal of Ecology, , –. and Rural Livelihoods (eds B. Dickson, J. Hutton & W.M. Adams),
C E N T R A L S TAT I S T I C S O F F I C E O F B O T S WA N A (CSOB) () Poverty pp. –. Wiley-Blackwell, Oxford, UK.
of Botswana. Http://botswana.opendataforafrica.org/POOB/ M AC D O N A L D , D.W. & L OV E R I D G E , A.J. () The Biology and
poverty-of-botswana-?village=-pandamatenga [accessed Conservation of Wild Felids. Oxford University Press, Oxford, UK.
June ]. M AC D O N A L D , D.W. & S I L L E R O -Z U B I R I , C. () Large carnivores
C LOV E R , J. & E R I K S E N , S. () The effects of land tenure change on and conflict: lion conservation in context. In Lion Conservation
sustainability: human security and environmental change in Research. Workshop : Modelling Conflict (eds A.J. Loveridge,
Southern African savannas. Environmental Science & Policy, T. Lynam & D.W. Macdonald), pp. –. Wildlife Conservation
, –. Research Unit, Oxford University, Oxford, UK.
D I C K M A N , A.J. () Complexities of conflict: the importance of M A D D E N , F. & M C Q U I N N , B. () Conservation’s blind spot: the
considering social factors for effectively resolving human–wildlife case for conflict transformation in wildlife conservation. Biological
conflict. Animal Conservation, , –. Conservation, , –.
D I C K M A N , A.J. & H A Z Z A H , L. () Money, myths and man-eaters: M E Y E R , L.C.R., F I C K , L., M AT T H E E , A., M I T C H E L L , D. & F U L L E R , A.
complexities of human–wildlife conflict. In Problematic Wildlife: () Hyperthermia in captured impala (Aepyceros melampus):
a Cross-Disciplinary Approach (ed. F.M. Angelici), pp. –. a fright not flight response. Journal of Wildlife Diseases, , –.
Springer, Cham, Switzerland. N A I D O O , R., C H A S E , M.J., B E Y T E L L , P., D U P R E E Z , P., L A N D E N , K.,
D I C K M A N , A.J., H I N K S , A.E., M AC D O N A L D , E.A., B U R N H A M , D. & S T U A R T -H I L L , G. & T AY LO R , R. () A newly discovered wildlife
M AC D O N A L D , D.W. () Priorities for global felid conservation. migration in Namibia and Botswana is the longest in Africa. Oryx,
Conservation Biology, , –. , –.
E S T E S , J.A., T E R B O R G H , J., B R A S H A R E S , J.S., P O W E R , M.E., B E R G E R , J., NOAA (N AT I O N A L O C E A N I C A N D A T M O S P H E R I C
B O N D , W.J. et al. () Trophic downgrading of Planet Earth. A D M I N I S T R A T I O N ) () Climate Data Online Search.
Science, , -. Http://www.ncdc.noaa.gov/cdo-web/search [accessed July ].
H A R R I S , G., T H I R G O O D , S., H O P C R A F T , J.G.C., C R O M S I G T , J.P.G.M. & O R I O L -C O T T E R I L L , A., M A C D O N A L D , D.W., V A L E I X , M., E K WA N G A ,
B E R G E R , J. () Global decline in aggregated migrations of large S. & F R A N K , L.G. () Spatiotemporal patterns of lion space use in
terrestrial mammals. Endangered Species Research, , –. a human-dominated landscape. Animal Behaviour, , –.
H A Y WA R D , M.W. & S LO T OW , R. () Temporal partitioning of P AC K E R , C., S WA N S O N , A., I K A N D A , D. & K U S H N I R , H. () Fear of
activity in large African carnivores: tests of multiple hypotheses. darkness, the full moon and the nocturnal ecology of African lions.
South African Journal of Wildlife Research, , –. PLOS ONE, , e.
H E M S O N , G. () The ecology and conservation of lions: human– P AT T E R S O N , B.D., K A S I K I , S.M., S E L E M P O , E. & K AY S , R.W. ()
wildlife conflict in semi-arid Botswana. PhD thesis. University of Livestock predation by lions (Panthera leo) and other carnivores on
Oxford, Oxford, UK. ranches neighboring Tsavo National Parks, Kenya. Biological
H E M S O N , G., M A C L E N N A N , S., M I L L S , G., J O H N S O N , P. & Conservation, , –.
M AC D O N A L D , D. () Community, lions, livestock and money: P O L I S A R , J., M A X I T , I., S C O G N A M I L LO , D., F A R R E L L , L., S U N Q U I S T ,
a spatial and social analysis of attitudes to wildlife and the M.E. & E I S E N B E R G , J.F. () Jaguars, pumas, their prey base, and
conservation value of tourism in a human–carnivore conflict in cattle ranching: ecological interpretations of a management
Botswana. Biological Conservation, , –. problem. Biological Conservation, , –.
H E T E M , R.S., M I T C H E L L , D., D E W I T T , B.A., F I C K , L.G., M E Y E R , P O U G H , F.H., J A N I S , C.M. & H E I S E R , J.B. () Vertebrate Life. th
L.C.R., M A LO N E Y , S.K. & F U L L E R , A. () Cheetah do not edition. Pearson, London, UK.
abandon hunts because they overheat. Biology Letters, , . R D E V E LO P M E N T C O R E T E A M () R: A Language and Environment
I N S K I P , C. & Z I M M E R M A N N , A. () Human–felid conflict: a review for Statistical Computing. R Foundation for Statistical Computing,
of patterns and priorities worldwide. Oryx, , -. Vienna, Austria. Https://www.r-project.org/ [accessed October ].
K A N S K Y , R., K I D D , M. & K N I G H T , A.T. () Meta-analysis of R E D P AT H , S.M., G U T I É R R E Z , R.J., W O O D , K.A. & Y O U N G , J.C. (a)
attitudes toward damage-causing mammalian wildlife. Conservation Conflicts in Conservation: Navigating towards Solutions. Cambridge
Biology, , –. University Press, Cambridge, UK.
K A N S K Y , R., K I D D , M. & K N I G H T , A.T. () A wildlife tolerance R E D P AT H , S.M., B H AT I A , S. & Y O U N G , J. (b) Tilting at wildlife:
model and case study for understanding human wildlife conflicts. reconsidering human–wildlife conflict. Oryx, , –.
Biological Conservation, , –. R I P P L E , W.J., E S T E S , J.A., B E S C H T A , R.L., W I L M E R S , C.C., R I T C H I E ,
K E L L E R T , S.R., B L A C K , M., R U S H , C.R. & B A T H , A.J. () Human E.G., H E B B L E W H I T E , M. et al. () Status and ecological effects of
culture and large carnivore conservation in North America. the world’s largest carnivores. Science, , .
Conservation Biology, , –. S C H A L L E R , G.B. () The Serengeti Lion: a Study of Predator–Prey
K G AT H I , D.L., M M O P E LWA , G., M A S H A B E , B. & M O S E P E L E , K. () Relations. University of Chicago Press, Chicago, USA.
Livestock predation, household adaptation and compensation S C H I E S S -M E I E R , M., R A M S A U E R , S., G A B A N A P E LO , T. & K Ö N I G , B.
policy: a case study of Shorobe village in northern Botswana. () Livestock predation—insights from problem animal
Agrekon, , –. control registers in Botswana. The Journal of Wildlife Management,
K U I P E R , T.R., L O V E R I D G E , A.J., P A R K E R , D.M., J O H N S O N , P.J., H U N T , , –.
J.E. & S TA P E L K A M P , B. et al. () Seasonal herding practices S O U L S B U R Y , C.D. & W H I T E , P.C.L. () Human–wildlife
influence predation on domestic stock by African lions along a interactions in urban areas: a review of conflicts, benefits and
protected area boundary. Biological Conservation, , –. opportunities. Wildlife Research, , –.
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217Environmental predictors of livestock predation 657
S TA H L , P., V A N D E L , J.M., H E R R E N S C H M I D T , V. & M I G O T , P. () USNO (U N I T E D S TAT E S N AVA L O B S E R VA T O R Y ) () Fraction
Predation on livestock by an expanding reintroduced lynx of the Moon Illuminated. Http://aa.usno.navy.mil/data/docs/
population: long-term trend and spatial variability. Journal of MoonFraction.php [accessed July ].
Applied Ecology, , –. V A L E I X , M., H E M S O N , G., L OV E R I D G E , A.J., M I L L S , G. &
S T R Y K E R , J.A. () Thermoregulatory behaviour assessment and M AC D O N A L D , D.W. () Behavioural adjustments of a large
thermal imaging of large felids. PhD thesis. The University of carnivore to access secondary prey in a human-dominated
Guelph, Guelph, Canada. landscape. Journal of Applied Ecology, , –.
S U N Q U I S T , M.E. & S U N Q U I S T , F.C. () Ecological constraints on V A N B O M M E L , L., B I J D E V A AT E , M.D., D E B O E R , W.F. & D E I O N G H ,
predation by large felids. In Carnivore Behavior, Ecology, and H.H. () Factors affecting livestock predation by lions in
Evolution (ed. J. Gittleman), pp. –. Springer, Dordrecht, Cameroon. African Journal of Ecology, , –.
The Netherlands. W A N G , S.W. & M AC D O N A L D , D.W. () Livestock predation by
T O R TA T O , F.R., L AY M E , V.M.G., C R AW S H AW , P.G. & I Z Z O , T.J. () carnivores in Jigme Singye Wangchuck National Park, Bhutan,
The impact of herd composition and foraging area on livestock Biological Conservation, , –.
predation by big cats in the Pantanal of Brazil. Animal Conservation, W E S T , P.M. & P AC K E R , C. () Sexual selection, temperature,
, –. and the lion’s mane. Science, , –.
T R E V E S , A., N A U G H T O N -T R E V E S , L., H A R P E R , E.K., M L A D E N O F F , W I L L M E R , P., S T O N E , G. & J O H N S T O N , I. () Environmental
D.J., R O S E , R.A., S I C K L E Y , T.A. & W Y D E V E N , A.P. () Predicting Physiology of Animals. nd edition. Wiley-Blackwell, New Jersey, USA.
human–carnivore conflict: a spatial model derived from years of Y U , L, E N D I , Z, Z H I H O N G , L. & X I AO J I E , C. () Amur tiger
data on wolf predation on livestock. Conservation Biology, , (Panthera tigris altaica) predation on livestock in Hunchun Nature
–. Reserve, Jilin, China. Acta Theriologica Sinica, , –.
Oryx, 2020, 54(5), 648–657 © 2019 Fauna & Flora International doi:10.1017/S0030605318001217
Downloaded from https://www.cambridge.org/core. IP address: 46.4.80.155, on 23 Nov 2021 at 22:52:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0030605318001217You can also read
