The observational work was carried out in two South African protected areas, Addo Elephant National Park and Marakele National Park, where the team accumulated more than 90 hours of direct observation of elephants in competitive, or agonistic, contexts. From those hours the researchers logged approximately 840 individual gesture instances, each one catalogued by form, modality, and the apparent goal it served. The dataset allowed the team to ask a question that has long fascinated cognitive scientists: when an elephant flings its trunk or spreads its ears wide at a rival, is it simply reacting, or is it trying to change the behavior of another individual on purpose?

The answer, the researchers report, is that the elephants were gesturing in a recipient-directed and goal-directed way, which is precisely what theories of intentional communication predict. Before and during a gesture, elephants visually checked whether the intended recipient was paying attention to them. They adapted the modality of their signal according to the recipient’s visual attention, for example favoring visual gestures when the other animal was looking and shifting toward tactile signals when it was not. After gesturing, they waited for a response. And when the recipient failed to respond in a way that satisfied their apparent goal, they persisted, producing further gestures. This chain of behaviors, attention-checking, audience-adjusting, response-waiting, and persistence, is considered a hallmark of first-order intentionality in animal communication research.

Lead author Vesta Eleuteri described the phenomenon in vivid terms. If you spend time observing wild elephants, she noted, you cannot help but notice their conspicuous gestures: they may fling their trunk or spread their ears wide, and an observer intuitively understands that they mean business. But until now, she explained, nobody knew whether wild elephants were using these gestures to communicate intentionally. The new findings show, for the first time, that in competitive situations wild elephants intentionally use over 40 different gesture types, composed of 32 gesture actions, to achieve two main goals: asking their recipients to move away, or to stop what they are doing.

One of the most striking results concerns flexibility. Of the 33 gesture types recorded in successful communications, 21, or 64 percent, were used for both meanings. That means the same physical gesture, an ear flap, a trunk throw, or a directed approach, could serve either to push a rival back or to halt an ongoing action, depending on the context. Elephants, in other words, do not rigidly pair each gesture with a single fixed goal the way a fixed signal in a simple signaling system would. Instead, they communicate flexibly, relying on shared context and the recipient’s response to disambiguate meaning. This kind of flexible, context-dependent repertoire had previously been documented in apes and a handful of other primates, making its appearance in elephants a remarkable case of evolutionary convergence.

Eleuteri emphasized exactly that point. Such extensive, flexible gestural repertoires have previously been shown in apes and a few other primates, she said, and finding the same in elephants is particularly exciting because elephants are so evolutionarily distant from primates. The two lineages last shared a common ancestor on the order of a hundred million years ago, so the similarity in communicative sophistication cannot be explained by close relatedness. Instead, the researchers suggest, sophisticated gestural communication may have evolved convergently in species that share advanced intelligence and complex social lives, a pattern that also raises questions about which other big-brained, long-lived social animals might possess comparable repertoires that have simply never been systematically catalogued.

The study also took a comparative step across populations. The team compared the gestures used by the two South African populations they observed, and then compared their findings with the ElephantVoices database, which is based on long-term observations of elephant populations in Kenya and Mozambique. The result was notable: the populations shared the same gesture types. That overlap suggests elephants may share a common gestural vocabulary across populations, rather than each group inventing its own idiosyncratic signals. For a species with fission-fusion social dynamics, in which family groups split and recombine over vast ranges, a broadly shared signal set would make it easier for elephants that have never met to understand one another quickly, including in tense competitive encounters.

Co-author Dr Lucy Bates of the University of Portsmouth highlighted the sheer effort behind the discovery. Tracking and understanding that communication is not easy, she said, because it takes a lot of painstaking work. Identifying so many different gesture types, as ways of communicating clear messages to move away or stop what you are doing, is really exciting, she added. She drew a direct parallel to humans, who use gesture all the time to get meaning across to the people they are talking with. Knowing that elephants also use these signals to communicate with each other, she suggested, points to gesturing as a possible feature of highly intelligent animals that lead complex social lives, rather than a quirk of any single lineage.

Senior author Professor Angela Stoeger of the University of Vienna, whose primary research focus is elephant vocal communication, placed the findings in a broader multimodal context. The results, she said, underscore why studying multiple modalities is essential. In the future, she explained, studying the interplay of acoustic, visual, tactile, and seismic signals, together with context, will finally reveal the full messages elephants exchange. Elephants are famous for producing low-frequency rumbling calls that travel through the air and ground, and they also touch one another constantly with their trunks. The new gestural dataset provides a foundation for understanding intentional visual signaling in wild elephants, but the authors stress that it is only a starting point, and they call for further research into how elephants use gestures in other contexts beyond competition, such as cooperation, caregiving, and courtship.

The findings also carry practical weight for conservation. Elephants face a range of pressures, including habitat loss that increasingly brings them into conflict with humans across their shrinking ranges. Dr Bates argued that the work matters precisely because humans and elephants increasingly come into contact as human populations expand in elephant range states. Understanding how elephants communicate when they are feeling threatened, she said, could help people better read their intentions or recognize when they do not want humans around, giving people the chance to move away before interactions become dangerous. In effect, the same gesture vocabulary that elephants use to tell a rival to back off could one day serve as an early-warning system for people living at the edge of elephant country, translating a flared ear or a thrust trunk into a clear and potentially life-saving instruction.

Subject of Research: Intentional gestural communication in wild African savannah elephants during competitive interactions

Article Title: ‘Move away’ and ‘stop that!’: Wild elephants use a rich gestural repertoire intentionally to “mean business”

Article References: ‘Move away’ and ‘stop that!’: Wild elephants use a rich gestural repertoire intentionally to “mean business”. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: elephants, animal communication, gestures, intentionality, Animal Cognition, African savannah elephant, behavioral ecology, evolutionary convergence, conservation, human-elephant conflict, South Africa, Move

The research team, led by scientists affiliated with Hawassa University and Addis Ababa University in Ethiopia, compiled more than 6,800 occurrence records for Loxodonta africana from field surveys in Ethiopian national parks, published literature, and the Global Biodiversity Information Facility. After removing duplicate and spatially redundant records, 4,298 verified presence points remained, spanning eight East African countries: Eritrea, Ethiopia, South Sudan, Somalia, Kenya, Uganda, Tanzania, and Rwanda. The authors restricted the dataset to records documented since the 1990s to avoid contaminating the models with historical distributions that no longer reflect where elephants actually live.

To translate those sightings into a predictive map, the researchers built an ensemble species distribution model that combined seven algorithms: three regression-based methods, namely generalized linear models, generalized additive models, and multivariate adaptive regression splines, and four machine-learning approaches, including boosted regression trees, maximum entropy, random forests, and support vector machines. Each algorithm was trained on 70 percent of the occurrence data and validated on the remaining 30 percent using a ten-fold subsampling scheme, with 10,000 background points generated across the study area to represent pseudo-absences. The final ensemble prediction was produced by averaging the individual model outputs, weighting each by its true skill statistic, a standard technique for reducing the uncertainty inherent in any single algorithm.

The environmental backbone of the models came from twenty predictor variables: nineteen bioclimatic layers drawn from the WorldClim 2.1 database at roughly one-kilometer resolution, plus a human footprint index that quantifies cumulative pressure from roads, settlements, agriculture, and infrastructure. Because many climate variables are strongly correlated with one another, the team screened them for multicollinearity, retaining only predictors with pairwise correlations below 0.7 and variance inflation factors under 5. Ten variables survived the filtering. Future projections relied on the HadGEM3-GC global circulation model from the Coupled Model Intercomparison Project Phase 6, run under two shared socioeconomic pathways: SSP2-4.5, an intermediate emissions scenario, and SSP5-8.5, a very high emissions pathway in which atmospheric carbon dioxide roughly triples by 2100.

The models performed exceptionally well. The average area under the receiver operating characteristic curve reached 0.94, with random forests and support vector machines delivering the strongest individual results, and the machine-learning algorithms significantly outperformed the regression-based approaches. Sensitivity and specificity values of 0.91 and 0.86 respectively indicated that the models were both reliable at identifying where elephants can live and where they cannot. Importantly, the predicted current suitable habitat of approximately 887,000 square kilometers under the weighted-mean threshold closely matched the species’ extant range of about 889,000 square kilometers delineated by the IUCN, lending the projections considerable credibility.

Which environmental factors mattered most? Not the human footprint, surprisingly. At the broad regional scale of the analysis, the dominant drivers of elephant habitat suitability were climatic: precipitation of the warmest quarter, isothermality, the mean temperature of the driest quarter, and precipitation of the driest month. These variables govern water availability, forage productivity, and thermal stress tolerance, all of which directly shape elephant distribution, migration corridors, and survival. The authors note that while climate determines the overall pattern of suitable habitat across East Africa, human pressures such as agricultural expansion, roads, and poaching remain critical local threats that fragment landscapes and squeeze the corridors elephants need to move between resource patches.

The future projections are where the study turns alarming. Under the intermediate SSP2-4.5 scenario, mean suitable habitat is projected to decline by 51.9 percent by 2050 and 52.6 percent by 2070 compared with the current average of roughly 821,000 square kilometers. Under the worst-case combinations, losses reach as high as 71.3 percent by 2050 and 73 percent by 2070, depending on the threshold applied. New habitat gains are minimal, amounting to only about 4 to 5 percent of the current range, meaning the species faces a net contraction of nearly half its climatic niche. The Horn of Africa fares worst of all: suitable areas in South Sudan and Somalia are projected to become unsuitable entirely, and the already fragmented habitats of Eritrea and eastern Ethiopia shrink to isolated patches.

Perhaps the most sobering finding concerns protected areas. East Africa’s protected area network covers about 573,000 square kilometers, yet only 17.2 to 17.9 percent of the currently suitable elephant habitat falls inside it. More than 82 percent of predicted suitable habitat lies outside formal protection, exposed to agricultural conversion, charcoal production, and infrastructure development. The concept of Anthropocene refugia, areas that remain climatically suitable and protected over time, offers a framework for identifying the landscapes most likely to sustain elephants through the coming decades. The study found that only about 21 percent of the remaining stable habitat projected for 2050 and 2070 sits within protected areas, and these refugia themselves decline under the harsher emissions scenario, shrinking by more than 21,000 square kilometers between the moderate and severe 2050 projections.

The authors caution that legal designation alone does not guarantee viable habitat. Substantial portions of the existing protected area network are already climatically unsuitable for elephants, and many reserves face severe encroachment and anthropogenic pressure. Elephants surviving in marginal or degraded habitats may function as refugee species, confined to forests and fragments to avoid people rather than occupying the landscapes their climatic tolerances would predict. This behavioral compression can mask the true extent of range loss and complicates conservation planning, particularly in the Horn of Africa where continuous population monitoring is limited and corridor establishment remains insufficient.

The study’s conclusions point toward an urgent agenda: reassess and realign protected area boundaries so that future climatic refugia are actually captured within conservation networks, establish transboundary corridors to maintain connectivity across national borders, strengthen community engagement and stewardship, and protect current habitats even where models forecast future unsuitability, especially for the vulnerable populations of Eritrea, Somalia, and South Sudan. Neither the suitable habitats inside protected areas nor those outside them, the authors conclude, can by themselves guarantee the long-term survival of the species. With Africa’s savanna elephants already down 30 percent over recent decades according to continent-wide surveys, and roughly 415,000 individuals of both African elephant species remaining, the window for proactive, climate-informed conservation in East Africa is narrowing rapidly. Mapping where elephants can persist, and securing those places before the climate closes them off, may be the difference between a managed decline and a functional extinction across much of the species’ eastern range.

Subject of Research: Climate-driven habitat suitability and Anthropocene refugia for the African savanna elephant in East Africa

Article Title: Persistence Vulnerability of the African Savanna Elephant Loxodonta africana to Anthropocene Threats in East Africa

Article References: Ahmed, A. S., Melese, D., Aligaz, M. A., Atickm, A., & Kufa, C. A. (2026). Persistence Vulnerability of the African Savanna Elephant Loxodonta africana to Anthropocene Threats in East Africa. Ecology and Evolution, 16(9), Article e74389. https://doi.org/10.1002/ece3.74389

Image Credits: AI Generated

DOI: 10.1002/ece3.74389

Keywords: African savanna elephant, Loxodonta africana, species distribution modeling, climate change, Anthropocene refugia, East Africa, protected areas, habitat loss, ensemble models, human-elephant conflict, conservation planning, Horn of Africa

The study, published in the journal Environmental and Sustainability Indicators, represents a milestone for conservation science in a region where elephant corridors had never before been scientifically delineated. Between 2021 and 2025, the researchers conducted systematic field surveys across all elephant-inhabited districts of Nagaland, georeferencing direct and indirect evidence of elephant presence—dung piles, footprints, feeding marks, trails, sightings and crop-raiding paths—using GPS units, and complementing these observations with structured interviews of Forest Department staff and community members whose families have coexisted with elephants for generations.

Of the eleven corridors identified, nine remain functionally active, while two—the Baghty Valley–Nambor Wildlife Sanctuary corridor and the Sitap Range–Geleki Reserve Forest corridor—have been classified as inactive, with no confirmed elephant movement since the late 1980s and 2015, respectively. The functioning corridors vary enormously in their ecological character. The Intangki National Park–Dhansiri Reserve Forest corridor and the Singphan Wildlife Sanctuary–Abhyapur Reserve Forest corridor support resident elephant populations that move year-round, whereas others serve primarily as seasonal migratory routes, with peak use coinciding with the monsoon (May to August) and winter (October to January) crop-maturation periods. The number of elephants using individual corridors ranges from as few as ten seasonal migrants to between 80 and 100 animals traversing the Longtho Range–Changdang Beat–Hollongpar Gibbon Sanctuary route.

What makes the study particularly compelling is its methodological rigor. To characterize land use around each corridor, the team established a two-kilometer buffer and analyzed the 2024 Land Use/Land Cover dataset derived from high-resolution LISS-IV satellite imagery processed through supervised Maximum Likelihood Classification. To track vegetation change over time, they computed the Normalised Difference Vegetation Index, or NDVI—a remote sensing metric that measures photosynthetically active biomass by comparing near-infrared and red light reflectance—using Landsat 4-5 imagery for 1994 and 2009 and Landsat 8-9 imagery for 2024. Multi-temporal change detection was then performed for three periods: 1994–2009, 2009–2024, and the full 1994–2024 span.

The NDVI results tell a deceptively optimistic story. Between 1994 and 2009, approximately 65 percent of the study area experienced declining vegetation greenness, with 11.8 percent showing strong decline and 53.6 percent moderate decline. But between 2009 and 2024, the trend reversed dramatically: about 53.9 percent of the area showed moderate increases in greenness and 36.7 percent strong increases. Taken at face value, the landscape appears to be recovering. Field verification, however, revealed otherwise. Much of the observed greening corresponds to the expansion and maturation of commercial plantations—rubber, areca nut, teak, tea and agar—rather than the restoration of structurally intact native forests. NDVI, the authors caution, cannot distinguish between a biodiverse tropical forest and a monoculture rubber estate, even though the two offer vastly different ecological value to elephants.

This distinction matters profoundly. Asian elephants are large-bodied mammals with substantial daily foraging requirements, depending on structurally complex forests that provide diverse browse species, seasonal forage and movement cover. Traditional jhum cultivation—the shifting agriculture practiced for centuries by Naga communities—maintained heterogeneous landscape mosaics with fallow cycles that allowed secondary forest regeneration and preserved habitat permeability. The conversion of jhum lands into permanent commercial plantations has fundamentally transformed vegetation structure, diminishing both palatable forage and habitat heterogeneity. Many corridors now consist of structurally simplified, plantation-dominated forests that, despite retaining vegetative cover, offer reduced ecological value for elephant movement.

The consequences of connectivity loss are perhaps starkest in the Baghty Valley. Historically, between 100 and 130 elephants moved through this corridor connecting Wokha district with Nambor Wildlife Sanctuary. Encroachment of Rengma Reserve Forest, rural and urban expansion, high-tension power lines, village roads and small industries severed the route entirely, isolating what is now a resident population of approximately 180 elephants. Trapped within a constrained landscape, these animals have expanded their range into human settlements, generating the region’s most severe human-elephant conflict—affecting 90 villages in Wokha, 11 in Zunheboto, five in Tseminyu and 15 in Mokokchung. The conflict toll there includes 12 human casualties and 21 unofficially recorded elephant deaths over three decades.

Community interviews conducted across the corridor landscapes revealed that the true scale of elephant mortality far exceeds official records. Most deaths were attributed to retaliatory shootings following repeated crop depredation and property damage, with fewer incidents from electrocution and accidental drowning. Because such events often went unreported, official statistics dramatically underestimate both the intensity of conflict and the number of elephants killed. Crop damage emerged as the most prevalent conflict type throughout the network, followed by destruction of plantations and homes, with paddy, rubber, areca nut and banana among the most frequently damaged assets.

Yet the study is not simply a chronicle of decline. In a finding that surprised the researchers, two corridors—Kuhuboto–Bokajan and Doyapur–Daldali—experienced renewed elephant movement in 2023 and 2024 after prolonged inactivity, demonstrating that partially degraded corridors can recover when sufficient structural connectivity persists. Local ecological knowledge offers a possible explanation: elephants appear to maintain culturally transmitted “memory paths” across generations. As one captive elephant owner from the Tuli range told the researchers, “These elephants, like humans, pass down the knowledge… even if the elephants don’t come today, their next kin… will at least come once in the traditional route their forefather once roamed.”

The governance context of Nagaland adds a distinctive dimension to the conservation challenge. Under Article 371(A) of the Indian Constitution, approximately 88 percent of the state’s land is owned by communities, clans or individuals rather than the government. Statutory protection is therefore largely irrelevant to corridor conservation; everything depends on community stewardship. Encouragingly, village councils and community leaders expressed general willingness to designate degraded or conflict-prone forest patches as community conserved areas or reserves—provided the government delivers sustained support through livelihood assistance, timely compensation and long-term conservation incentives. Resistance to elephant conservation, the authors conclude, stems less from opposition to wildlife than from the socio-economic costs of coexistence.

The researchers propose corridor-specific interventions rather than uniform strategies. Recovering corridors such as Kuhuboto–Bokajan and Doyapur–Daldali should be proactively protected before further degradation occurs. Severely fragmented corridors, including Baghty Valley–Nambor and Sitap–Geleki, require active ecological restoration through assisted natural regeneration and enrichment planting of indigenous forage species. The planned reclamation of roughly 1,980 hectares of encroached forest within Rengma Reserve Forest offers a critical opportunity to restore transboundary connectivity for the isolated Baghty elephant population. Regulating coal mining, quarrying and other extractive industries within corridor buffers—currently concentrated in Mon, Longleng, Mokokchung and Wokha districts—constitutes another immediate priority.

Because every identified corridor crosses the Nagaland–Assam boundary, the authors emphasize that coordinated, transboundary governance is essential; conservation efforts implemented independently by either state are unlikely to succeed. Ultimately, the study reframes elephant corridors not as static lines on a map but as dynamic socio-ecological systems where ecological connectivity, indigenous governance and rural livelihoods are inseparably intertwined. With community-led conservation, equitable compensation and landscape-scale planning, Nagaland may yet establish a nationally significant model of decentralised corridor management—one that secures the future of its elephants while sustaining the people who share their forests.