Managing elephanthuman conflicts

Elephants that threaten human lives and agricultural crops have been dealt with historically by society in various ways, from swift retaliation to resigned coexistence. At either end, there is suffering—for elephants or for people. Indeed, the losses suffered due to elephants are only a part of the broader wildlife-human conflict, including depredation by many other animals (table 9.1). In oil palm plantations in Southeast Asia, only about half the damage can be attributed to elephants, while significant losses also occur due to wild pig and porcupine.

Although there are few objective estimates of the economic loss inflicted by elephants, it is certain that the losses run into several tens of millions of dollars each year across the two continents. During the 1970s, it was estimated that the Federal Land Development Agency in Malaysia alone lost oil palm and rubber plantation crops worth U.S.$20 million annually in the export market due to damage by elephants. I made a conservative estimate of U.S.$0.5 million loss to agricultural crops annually in southern India during the early 1980s. The southern state of Karnataka, home to about 20% of India's population of over 25,000 wild elephants, paid an average of U.S.$160,000 per year during 1991-2000 as compensation to farmers for crop loss due to elephants. The losses have been much higher in other parts of the country, such as the north-

Table 9.1

Damage to oil palm trees by various wild mammals in the FELDA plantations of peninsular Malaysia and TIGAMITRA plantations in Aceh, Sumatra.

Table 9.1

Damage to oil palm trees by various wild mammals in the FELDA plantations of peninsular Malaysia and TIGAMITRA plantations in Aceh, Sumatra.

FELDA, Malaysia*


No. of Trees

No. of Trees




Damaged %




119,068 45.9

Wild pig



65,305 25.2




30,073 11.6


17,201 6.6


27,827 10.7

Sources: Blair et al. (1979) for FELDA plantations and Sukumar (1999, unpublished data) for

TIGAMITRA plantations. *The data for FELDA are up to 1978 with most of the damage occurring during 1976-1978 though plantations were opened up since 1956. Some damage by rodents also occurs but a precise figure is not given in Blair et al. (1979). tThe data for TIGAMITRA are for the period 1995-1998.

Sources: Blair et al. (1979) for FELDA plantations and Sukumar (1999, unpublished data) for

TIGAMITRA plantations. *The data for FELDA are up to 1978 with most of the damage occurring during 1976-1978 though plantations were opened up since 1956. Some damage by rodents also occurs but a precise figure is not given in Blair et al. (1979). tThe data for TIGAMITRA are for the period 1995-1998.

east, where conflict has been more severe. The state of West Bengal, which has less than 500 elephants, paid an average annual compensation of U.S.$175,000 during 1996-2000 for crop losses due to wildlife, mostly elephant. In both states, the actual losses would have been higher.

At the same time, many elephants are killed by farmers as a direct consequence of agricultural conflict. Records I have maintained for southern India show that at least 230 elephants died in such conflict during 1976-2000, representing over 8% of all recorded elephant deaths.

Experienced and pragmatic observers of elephants in Asia and in Africa agree that elephant-human conflicts cannot be totally eliminated, except through the elimination of substantial numbers or proportions of elephants. The goal of management should thus be to control and minimize conflict. The available options to manage conflict can be considered under three broad categories: preventing or discouraging elephants from entering settlements and crop fields, elephant population management, and a suite of measures pertaining to social welfare, economic incentives, and land-use planning. A conflict management scheme must obviously be economically viable for it to be sustained in the long term. The Great Wall of China can certainly exclude elephants, but that does not make economic sense. The challenge, therefore, lies in formulating region-specific measures that are both effective in containing conflict and based on sound benefit-cost considerations.

Traditionally, rural societies in Asia and Africa have sought protection from wild animals by organizing themselves into tight-knit communities and cooperating to keep animals away from their settlements and fields. To deal with elephants, farmers have relied on simple means, such as shouting, noise-making with pots and pans, drums, or firecrackers, and hurling rocks or fireballs at them. These are of limited use in keeping elephants at bay. Relatively inexperienced crop raiders may be scared away by such tactics. Veteran raiders, usually adult bulls or even some family groups, are not fooled. When possible, farmers guard their fields from the safety of a treetop platform. Those who try to scare elephants from flimsy structures on the ground are vulnerable to aggressive elephants. A bright spotlight powered by a car battery seems to be effective, but a weak torchlight may only invite a charge from an aggressive bull. Farmers who use dogs to guard their crops may also invite the wrath of raiding elephants.

The northern Bengal region in India experiences very intense conflict, with elephants not only damaging crops, but also demolishing dwellings and killing people. Squads of trained guards and villagers armed with firecrackers and guns, sometimes assisted by trained elephants, have been moderately successful at chasing away marauding elephants in this region. Trenching along the forest-cultivation boundary was attempted around oil palm plantations in peninsular Malaysia during the 1970s, but was soon abandoned because of rapid soil erosion caused by the high rainfall.

Some Indian states, such as Karnataka, have taken to trenching on a significant scale around protected areas (fig. 9.4). The reasoning is that such "barriers" should not only serve to keep elephants away from cultivation, but also prevent livestock from entering the forest. Trenches have been relatively ineffective in higher rainfall areas, such as the western boundary of Nagarahole National Park because of erosion and breaks occurring at places where streams cut across the trench. Trenching has been more effective in areas of low rainfall, such as the northeastern boundary of Bandipur National Park, where the soil is harder and the substrate rockier.

A trench has to be at least 2 m deep, 2 m across at the top, and 1.5 m across at the bottom to keep out an elephant. Even then, a determined elephant could fill up a trench partially by pushing in the soil with its feet and negotiating it. Depending on the nature of the soil, a trench with these minimum specifications costs between U.S.$2,000-$4,000 per kilometer to excavate. Measures to stabilize the wall of a ditch by planting on excavated soil heaped on the outer boundary are inexpensive. Masonry work along the wall is expensive. The considerable additional costs of plugging gullies and streams can result in an unfavorable benefit-cost ratio for a trench. Possible alteration of local drainage by a long trench should also be a consideration. Overall, a trench is most unlikely to be viable in regions with annual rainfall above 100120 cm, while in those with less rainfall, it could work when a hard substrate keeps maintenance costs low.

A boulder wall held together by wire mesh is an effective barrier against elephants. Its use is limited to areas where boulders are available and transporting costs are low. Dry stone walls in Kenya's Laikipia district have been generally ineffective because elephants push these over with their chests. Barriers

Figure 9.4

Ditches and electric fences are two commonly used barriers to keep elephants away from agricultural land. At the top is a reasonably effective trench in dry, hard substrate in Bandipur National Park, India. The bottom shows an official (left) of the Kenya Wildlife Service explaining the features of an electric fence to the author (right).

Figure 9.4

Ditches and electric fences are two commonly used barriers to keep elephants away from agricultural land. At the top is a reasonably effective trench in dry, hard substrate in Bandipur National Park, India. The bottom shows an official (left) of the Kenya Wildlife Service explaining the features of an electric fence to the author (right).

such as iron spikes embedded in concrete on the ground can be dangerous to elephants.

The high-voltage electrified fence is the most widely used barrier in the two continents (fig. 9.4). The basic principle behind this barrier is to deliver a strong shock that deters an animal, but is neither fatal nor injurious in any manner. This is achieved through the use of a transformer (the energizer), typically powered by a 12-volt car battery, that delivers a current of 5,00010,000 volts in pulses of very short duration (say, 1/3,000 second) with a gap of about 1 second between successive pulses. A minimum of 7,000 volts is recommended; several observers also point out that a fence in a moist region is more likely to provide sufficient grounding to deliver a strong shock. The battery itself is charged from the mains (110-230 volts alternating current [AC]) or by a solar panel in places where power supply through transmission lines is not available. The fence design itself may vary from a simple singe wire or 2-strand wire attached to existing trees to a sophisticated 12-strand barrier with high-quality accessories such as protected steel posts and insulators.

The effectiveness of an electric fence depends not only on fence design, pulse voltage, and maintenance, but also on the learning capacity and behavioral response of crop-raiding elephants. A fence may be broken by an elephant using its tusks (which are nonconductors) to prise an insulator or even break the wire. Bulls are also known to use the soles of the front feet, again poor conductors, to press down on fence wires. Fence posts may be broken by kicking with the legs or pushing with the tusks, while those protected with live wires may also be dealt with by breaking the wire between posts. Elephants may push trees over a fence, and a determined one may even crash through a live wire without receiving a shock if the timing of the break is between two pulses.

A study of electric fences by C. R. Thouless and J. Sakwa in Kenya's Laikipia district showed no clear relationship between their effectiveness and factors such as sophistication of design and voltage. Some simple fences or those providing a current with only 3,000-4,000 volts were effective over long periods, while some high-specification fences, including a 12-strand one, were repeatedly breached by elephants.

A survey in southern India indicated that sound design, although not necessarily an elaborate one, and good maintenance of electric fences were important for their effectiveness. The ownership of an electric fence was also a key factor in its success because this had a direct bearing on its maintenance. Of 19 fences under private ownership, 16 inspected were functional, while only 3 of 18 fences put up by a government agency were working. The failure of the fences in the latter category was mainly because of breakage of wires by villagers to gain access for themselves and their livestock to the forest and the inability of the forest department to bear the high expenditure for maintenance under the circumstances. The two categories of fence ownership were basically of similar design specification. No elephant has passed through one fence in the Nilgiris for over 5 years because its owner continually incorporated innovative measures, thus keeping one step ahead of elephants. The high-voltage electric fence is thus only a psychological bluff and not a physical barrier. The learning experience of elephants during the initial period of contact with fences may also decide their future response to various fence designs. Innovative measures from farmers are needed against the behavioral response of individual elephants in this "arms race."

What is the overall effectiveness of electric fences? Do these fences have favorable benefit-cost ratios? Malaysia was possibly the first country to use electric fences extensively to protect oil palm and rubber plantations from elephants and other herbivores. A success rate of 80% against elephants has been reported. During early experiments in 1982 by Robert Piesse with a fence in Etosha, Namibia, a total of 259 elephants made 184 contacts without a single break. Caitlin O'Connell-Rodwell and associates recorded no claims against elephant damage in the village of Lianshulu, Namibia, during 1994 after installation of an electric fence, while there had been about 30 claims over the previous 2 years. They attributed this partly to the relative paucity of bull elephants there; the family groups were more hesitant to risk breaking through the fence. The study at Laikipia, Kenya, also concluded that family groups without a bull were much more likely to be deterred, compared to groups with a bull, bull groups, or solitary bulls, by an electric fence (table 9.2).

The cost of an electric fence can vary from about U.S.$500 per kilometer for a 3- or 4-strand fence using local materials up to U.S.$5,000 per kilometer for an elaborately designed 8-12-m strand fence using imported components. With a basic cost of U.S.$200-$1,000 for a good energizer unit, depending on its sophistication, the average costs per kilometer obviously decline with increasing length of a fence. Privately owned fences of simple design in southern India averaged about U.S.$500 per kilometer, while those installed by government agencies cost about twice this amount for a fence with similar specifications, partly because of labor costs.

Table 9.2

Number of elephant social groups crossing electric fence line outward from a ranch before and after erection of fence in Laikipia, Kenya.

Number of Groups Number of Individuals

Table 9.2

Number of elephant social groups crossing electric fence line outward from a ranch before and after erection of fence in Laikipia, Kenya.

Number of Groups Number of Individuals

Elephant Social Group





Single bull





Bull group










Cow/calf group





Source: Simplified from Thouless and Sakwa (1995). Reproduced with the permission of Elsevier Science.

Source: Simplified from Thouless and Sakwa (1995). Reproduced with the permission of Elsevier Science.

The early Malaysian fences cost U.S.$1,200 per kilometer. Fence costs are reported as U.S.$500-$1,500 per kilometer in Zimbabwe, while costs in Kenya typically are about U.S.$2,500 kilometer or higher because of the need to import several components. A well-designed, experimental, 2-km fence recently put up in Wyanad in southern India cost U.S.$3,700 per kilometer, with partial labor being provided by the villagers. To the above figures 10%-20% can be added as annual costs of maintenance.

There are few published benefit-cost analyses of electric fences. Based on data on fence costs, efficacy, and damage to oil palm plantations in peninsular Malaysia provided by James Blair and associates during the late 1970s, I computed that every dollar invested in fencing would save $74 over a 5-year period. Such a favorable benefit-cost ratio would apply only to commercial crops.

For Hasanur, a seriously affected village in southern India that I studied during 1981-1982, I further computed that the costs of a two-strand fence could be recovered through savings in crop yields within 2 years, and that the annual losses were about three times the annual costs of fence maintenance. The electric fence at Lianshulu, Namibia, costing U.S.$5,900, resulted in an annual saving of crops worth about U.S.$900 during 1994 based on damage during the preceding 2 years.

We need to keep in mind that the favorable economics of an electric fence during initial years could change substantially once determined elephants learn to break through a relatively simple fence. Although electric fencing seems to have been the most successful barrier against elephants, the economics may not justify a high-end fence. Thouless and Sakwa are thus of the opinion that it may be better to build low-cost fences of simpler design and use these as "no-go" areas, rather than impenetrable barriers, combined with management of elephants that persist in breaking the fence. In parts of Africa, this management could include selective shooting of elephants, while in Asia, this has to be confined mostly to capture.

Given the elephant's highly developed sense of smell, there has been interest in identifying possible chemical deterrents. Some experiments in India with the urine of a large predator such as the tiger in scaring elephants have been inconclusive. The use of a deer repellent named HATE-C4 (developed by a German company) in southern Africa against elephants produced equivocal results. A similar repellent used in the high-rainfall areas of northeastern India failed completely.

The most promising results so far with chemical repellents seem to be a capsicum-based irritant developed and deployed by Ferrel Osborn in Zimbabwe. Capsicum (or capsaicin), the chemical ingredient derived from dry, ripe fruits of several species of the genus Capsicum (peppers and chilies), is a local irritant to sensory nerve endings, but has no lasting effects on mammals. Osborn conducted several experiments with a 10% oleoresin solution of capsicum that was propelled toward raiding elephants from an aerosol canister. The atomized resin floats in a cloud that remains effective for about 20 minutes after traveling a distance of 50-75 m during light wind conditions. The typical response of an elephant exposed to the spray was to freeze, exhale air audibly, test the air with its trunk, shake its head vigorously, and move away rapidly, often after roaring or trumpeting. The experiments were conducted with elephants in the wild at Hwange National Park and with raiding elephants in agricultural areas of the Gokwe Communal Lands. Positive responses were obtained in 19 of 22 tests conducted at Hwange and in 16 of 18 tests at Gokwe. Because of the difficulties in identifying elephants at night, it was unclear whether any of the elephants exposed to the spray returned at a later time to raid crops. While it has been demonstrated that capsicum possesses short-term repellency toward elephants, more experiments are needed to see if animals can be adversely conditioned to avoid crop fields by this method.

Other delivery systems, such as exploding grenades, packed with capsicum powder, may also be useful under field conditions. It has been suggested that the chemical signals of pheromones could be mimicked (see chapter 4) to lure or scare elephants away from crop fields. Preliminary experiments with temporal gland secretion by M. L. Gorman produced ambiguous results. Ongoing research by Bets Rasmussen on chemical repellents seems to hold some promise.

Apart from traditional noise-making, there have been few systematic attempts to use sounds associated with a certain meaning or that convey fear to elephants. A tiger's call played back through a loudspeaker on a large farm in southern India apparently seems to have deterred elephants. The Maasai people of East Africa have a tradition of young men spearing wild animals, including elephants, as part of pubertal initiation rites. Kadzo Kangwana carried out a series of playback experiments with sounds of Maasai-associated cattle bells and mooing. The elephants responded by raising their heads, smelling the air, and retreating from the direction of the sounds. Clearly, they associated these cattle sounds with the Maasai. Such traditional, fear-invoking practices of people could be adopted locally to reduce depredation. Some observations made recently in Kenya suggest that elephants may be repelled by swarms of the African bee (Apis mellifera scutellata). Beehives could thus be deployed by farmers around their farms to deter elephants from venturing inside (Vollrath and Douglas-Hamilton, 2002).

The language of infrasound may also hold promise as a deterrent. Caitlin O'Connell-Rodwell played back natural distress calls of elephants recorded at Etosha, Namibia, to family groups and bulls in the same park. In one set of three trials, the family groups tested fled from the site on all occasions, while in other trials, the herds responded aggressively by surrounding the hide from which the calls were broadcast. Bulls were not disturbed by the distress calls. Clearly, much more work needs to be done to see if effective but cheap deterrents based on sound can be evolved.

The impact of elephants on agriculture and people has been kept under check historically through the elimination of the offending elephants, through either capture or killing in Asia and by killing in Africa. Some of this would have been selective targeting of raiding elephants. Much of this form of elephant population management has undoubtedly been haphazard and wasteful.

The large-scale capture of elephants in ancient times or the slaughter of elephants during the colonial period, ostensibly with the goal of control of problem animals, was indiscriminate toward raiders and nonraiders. Conflict control today undoubtedly requires some management of elephant populations through removal of animals, but on a more scientific basis.

Elephant population management can be considered under two kinds of circumstances. The majority of studies relating to elephant-human conflicts in the two continents have shown that, in the relatively intact habitats with large populations, the subadult and adult male elephants show a much higher propensity to raid crops compared to the female-led groups (chapter 7). Also, in economic terms, a major part of the damage to crops may be attributed to these male elephants in populations in which adult sex ratios are close to natural, and not significantly female biased.

Therefore, I have suggested that, when minimal intervention is desired, the first management option to reduce conflict should be to remove identified, notorious bulls selectively. This would not only reduce crop depredation significantly in the short term and appease farmers, but also usually lower the incidence of manslaughter. From an ecological viewpoint, the demography of the population would not be comprised in a polygynous species; field data clearly show that moderately female-biased sex ratios do not lower the birth rate or the population growth rate. Elephant populations that number in the thousands or several hundreds and are not already heavily female biased can absorb the selective removal of some bulls.

In the Indian state of Karnataka, about 60 "rogue" bull elephants were captured during 1985-2000 from regions where "surplus" bulls were available without making any adverse impact on demography. Four of these died during capture or subsequent transportation, while the rest have been trained and held in captivity. During the initial operations, some of the bulls were released into a national park over 100 km away, but at least two of them returned to the original site of capture.

The option of conflict management through selective removal of bulls is much more limited or not available in most other regions of southern India where poaching of male elephants for ivory has resulted in highly female-biased sex ratios (conflict is incidentally much lower in these regions). This form of management through selective removal of bull elephants has undoubtedly been practiced widely in other parts of Asia and in Africa. Interestingly, the Arthasastra, an ancient Indian treatise, prescribed the capture of only tusked, postpubertal bull elephants, the most prudent way of managing an elephant population.

Richard Hoare now believes, however, that when an offending bull is removed, a new "problem animal" soon replaces it; thus, this form of management may have to continue indefinitely. Obviously, we do not fully understand the social dynamics of bull elephants in relation to their raiding behavior.

The most common argument against the selective capture of rogue bull elephants is that these are the best breeders, and that the genetic future of the population would be compromised. First, there is no evidence that crop-raid ing bulls are indeed the best breeders even if there is that potential; remember, such individuals would be at risk of injury and elimination at the hands of farmers. Second, there are no reasons to believe that rogue bull elephants have any inherent genetic superiority (higher natural longevity or better resistance to disease, for instance) except perhaps by virtue of their dominance over other bulls. If removed from the population, other bulls could easily replace them for breeding. A genetic argument for nonintervention is weak when applied to large elephant populations in serious conflict with people. Where elephant populations are smaller, in tens or a few hundred, interventionist management should be based on population viability analyses that incorporate both demographic and genetic considerations.

A different situation seems to arise in smaller elephant populations ranging over highly fragmented habitats, such as in central India, where family groups indulge in as much, if not more, raiding than the bulls. The larger home range requirements of family groups, compared to bulls, at least in the short term may be one reason for this increased conflict. If such small elephant groups are in serious conflict and nonviable in demographic terms, the only option would be their removal along with the removal of associated bulls from the area. Even if they have a distinct genetic identity, these elephants are unlikely to contribute to the variation of any larger population in view of their isolation.

In Africa, the only way to remove problem animals has been to shoot them; this is unlikely to change in the near future. In Asia, the option of capturing elephants is available in most countries, although the expertise in taming large bulls is restricted to a few regions, such as southern India and parts of Myanmar. In the case of large bulls with a history of aggressive behavior and manslaughter, their elimination through shooting may be the only practical option in other places. Elephants have been captured in recent times by a combination of traditional methods, such as using trained captive elephants (koonkies) to restrain wild elephants after these are sedated or immobilized with a suitable drug.

A decision whether the captured elephants should be retained in captivity would depend on the local expertise available for training them and the demand for captive elephants. The transfer of captured problematic bull elephants to new locations has generally been a failure in Asia. Such bulls have either traced their way back to their original home range area or, if this has been impossible, continued to be in conflict with people in the area where they were released. In one case, a relocated bull in northern Bengal walked over 100 km back to its original range.

Fewer attempts have been made in translocating problematic family groups because of logistical limitations in Asian countries in handling entire herds. The drive has been the most commonly used method with family herds. There are several examples of successfully driving elephant herds, including very large ones, over distances exceeding 50 km in Sri Lanka, Sumatra, and India, but in the absence of postrelocation monitoring, it is unclear whether any of these elephants came into conflict with people.

Translocation of elephants, either bulls or herds, must therefore be accom panied by monitoring through radiotelemetry to decide on the future course of management. It is necessary to continue experimenting with relocation of elephants, especially herds, to learn more about factors that could contribute to its success.

0 0

Post a comment