The research team, led by Anjar Dimara Sakti of Institut Teknologi Bandung, built a habitat suitability index that treats the dragon’s ecological needs as a family of linked indicators rather than a single statistical guess. Because Komodo dragons are age-structured predators, the model separately characterised adult and juvenile requirements. Adults hunt ungulates such as Timor deer, water buffalo and feral pigs across open savanna, while juveniles are arboreal, feeding on small lizards, large insects and snakes in taller trees where they can escape cannibalism by their own kind. The framework therefore modelled three indicators in parallel: the distribution of adult prey, the distribution of juvenile prey, and the ecological conditions where the lizards themselves have been recorded.

To train the habitat model, the researchers drew 144 verified occurrence records of Varanus komodoensis from the Global Biodiversity Information Facility, filtering the citizen-science data by species identity, geography and coordinate uncertainty below one kilometre, a threshold matched to the animal’s daily roaming distance. Because absence data do not exist for such a wide-ranging predator, the team generated pseudoabsence points using a multi-criteria analysis that scored the landscape for adult prey, juvenile prey and ecological suitability, retaining only the lowest-scoring locations as clearly unsuitable ground. Those 169 pseudoabsences, balanced against the presences, fed into a Maximum Entropy model alongside seventeen environmental predictors spanning topography, climate, land cover, tree cover, canopy height and satellite-derived surface temperature.

The results paint a precise picture of dragon geography. Roughly 36 percent of Komodo National Park, about 20,885 hectares, falls into the high or very high suitability classes, with the best habitat concentrated on the northern side of Rinca Island, across much of Padar, and in low-lying, flat savanna with moderate tree cover. The model achieved an area under the receiver operating characteristic curve of 0.937, outperforming earlier Komodo habitat studies that reached 0.776 and 0.842 with narrower predictor sets. Unsurprisingly for a ground-dwelling ectotherm, surface temperature and the transitional mosaic between open grassland and closed forest emerged as decisive variables: the lizards shuttle between sun and shade to regulate body temperature, and the same mosaic sustains the deer and pigs that anchor the food web.

Habitat alone, however, does not determine survival, and the study’s most consequential move was to treat wildfire susceptibility as an inseparable part of the conservation equation. Komodo National Park’s lowland savanna, dominated by drought-adapted grasses and scattered lontar palms, cures into highly flammable fuel during the long May-to-October dry season, and fires burned 170 hectares of savanna between 2013 and 2021. The team trained three tree-based machine-learning algorithms, classification and regression trees, gradient tree boosting and random forest, on 150 wildfire points extracted from Sentinel Hub together with 250 non-fire points, using predictors that included elevation, slope, wind speed, land surface temperature, Sentinel-2 vegetation and moisture indices, and Sentinel-1 radar backscatter.

Random forest proved the strongest individual learner, with a mean AUC of 0.885 across 100 repeated data splits, followed by gradient tree boosting at 0.863. The researchers were candid about the ensemble’s limits: averaging all three algorithms equally dragged performance down because the weakest learner contributed as much as the strongest, and a two-model ensemble of random forest and gradient boosting scored higher at 0.878. More importantly, the team tested how well the models transferred beyond their training geography. Under spatially blocked cross-validation and leave-one-fire-out validation, in which each of the three mapped burn polygons was withheld in turn, discrimination fell from 0.912 on random partitions to 0.799 and 0.735 respectively, a sobering reminder that apparent accuracy can inflate when fire records cluster in accessible terrain.

The wildfire model’s explanatory power came from interpretable machine learning. Permutation importance and SHAP analysis converged on a consistent story: susceptibility rises as vegetation greenness (NDVI) and radar backscatter decline, as the burned-area index NBR+ increases, and as wind speed and maximum land surface temperature climb. In plain terms, the flammable landscape is sparse, dry, low-biomass savanna under warm, windy conditions, precisely the coastal lowlands the dragons favour. Slope and elevation, by contrast, contributed little once the vegetation indices were included, suggesting that topography shapes fire risk mainly through the vegetation it supports. The model also exposed a gap between official accounts and satellite reality: three fires officially reported at 10 hectares each in 2018 and 2021 measured 102.93, 184.63 and 300.07 hectares in Sentinel-2 imagery, with one blaze suspected of being set by poachers to divert rangers.

Overlaying the two surfaces produced the study’s central deliverable, a bivariate conservation priority map. About 58 percent of the park sits in the lowest wildfire vulnerability class, but 11 percent, some 7,200 hectares, is highly vulnerable, concentrated along low-elevation coastal savanna. Critically, 2,800 hectares, five percent of the park, combine high habitat suitability with high fire risk, and another 3,500 hectares of moderate suitability face the same threat. When the researchers overlaid seven Komodo movement paths from historical field monitoring, most dragon ranges fell in low-vulnerability zones, but the paths of two individuals crossed medium-to-high fire exposure, and one path contained the single largest share of high-suitability, high-risk ground. The analysis also revealed a mismatch between ecology and management: the park’s core protection zone holds less high-quality, low-risk habitat than the adjacent jungle zone, prompting the authors to recommend extending core-zone boundaries into those safer, high-suitability patches.

The study is explicit about its own uncertainties. The wildfire samples derive from only three burn events rather than 150 independent fires, the canopy height layer dates from 2005, and adding human-accessibility predictors produced an apparent accuracy gain that collapsed under spatial validation, indicating that ignition sources and flammable terrain coincide too closely in this small park to separate statistically. The authors stress that the priority maps should be treated as decision-support hypotheses requiring field confirmation, not precise boundaries for allocating suppression resources. Yet the framework’s real significance may lie beyond Komodo. By coupling an ecologically informed, multi-indicator species model with a threat-susceptibility model and a formal zoning evaluation, all built from free satellite and biodiversity data, the approach offers a transferable template for protecting other range-restricted species in seasonally dry, fire-prone landscapes, from island reptiles to fragmented mammal populations across Southeast Asia.

For the dragons themselves, the practical prescriptions are concrete: early dry-season fuel management and firebreaks around movement corridors in the coastal savanna overlap, tighter ignition controls given the suspected human role in past fires, seasonal restrictions on tourism in the highest-risk zones, and periodic re-running of both model layers so that zoning evolves with new data. With climate change degrading vegetation quality, prey populations declining on multiple islands, and plans to push tourism on Rinca toward more than a million visitors a year, the world’s largest lizard now has something it has never had before: a data-driven map of exactly where its most valuable ground and its most dangerous threat intersect.

Subject of Research: Machine learning-based habitat suitability and wildfire susceptibility modelling for Komodo dragon conservation planning in Komodo National Park, Indonesia

Article Title: Machine learning-based Komodo dragon conservation planning using habitat suitability and wildfire risk

Article References: Sakti, A. D., Zakiar, M. R., Rosleine, D., Artaningh, F., Santoso, C., Pramudya, A. D., Candra, D. S., Wijaya, S., & Wikantika, K. (2026). Machine learning-based Komodo dragon conservation planning using habitat suitability and wildfire risk. Environmental and Sustainability Indicators, 32, Article 101540. https://doi.org/10.1016/j.indic.2026.101540

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101540

Keywords: Komodo dragon, machine learning, habitat suitability, wildfire susceptibility, species distribution model, MaxEnt, random forest, remote sensing, Komodo National Park, conservation planning, Sentinel-2, biodiversity

The research team, working with the conservation organisation Rewilding Apennines as part of the LIFE ESC360 project, assembled 69 spatially explicit wildcat occurrence records collected between 2009 and 2023. The records spanned the full length of the country, from Friuli-Venezia Giulia in the northeast to the slopes of Mount Etna in Sicily, and drew on three independent datasets: the national ‘Progetto Gatto Selvatico’ archive, new camera-trapping surveys in Abruzzo, and previously published records from Etna Regional Park. To ensure data reliability, records from the national project were classified under the SCALP scheme, a verification framework originally developed for large-carnivore monitoring, yielding 11 genetically or morphologically confirmed records and 24 supported by unambiguous camera-trap photographs of wild-type individuals.

With those occurrences in hand, the team built a species distribution model using MaxEnt 3.4.3, a maximum-entropy algorithm designed for presence-only data. Environmental predictors were deliberately kept parsimonious to avoid overfitting: eight land-cover variables derived from the CORINE Land Cover 2018 dataset, aggregated to reflect habitat types relevant to wildcats, plus average elevation from the SRTM digital elevation model. Because camera traps and carcass recoveries are not distributed evenly across the landscape, the researchers generated a kernel density surface of sampling effort and used it to bias the selection of 10,000 background points, so that the model would not simply relearn where humans had looked. Model tuning via the ENMeval package selected hinge features and a regularisation multiplier of 3, and the final model was validated against a distribution of 100 null-model iterations.

The results were strikingly clear. The model achieved an AUC of 0.81, well above the null-model median of 0.50, and two variables dominated the prediction: forest and grassland cover contributed 62.5 percent of the model’s explanatory power, with a permutation importance of 78.6 percent, while average elevation contributed a further 36.1 percent. The response curves showed that the probability of wildcat occurrence rose sharply with increasing forest and grassland cover, levelling off at roughly 70 to 75 percent cover, and climbed with elevation up to about 1,000 metres above sea level before plateauing. Urban fabric, roads, agriculture, and water bodies contributed essentially nothing to the model. Mapped across Italy, the highest suitability values traced the spine of the Apennine chain and the Alpine region, with lower but non-negligible suitability in lowland and coastal zones.

To test whether these predictions held up on the ground, the researchers collected 23 new wildcat occurrences in the Central–Southern Apennines, none of which had been used to calibrate the model. Sixteen tissue samples came from carcasses, mostly roadkill and a few accidental hunting bycatch, recovered by rangers in the Abruzzo, Lazio and Molise National Park and the Regional Natural Park of the Simbruini Mountains. Seven further occurrences came from non-invasive monitoring combining camera traps with hair traps, wooden poles wrapped in Velcro and laced with valerian tincture, deployed in the Monte Genzana–Alto Gizio Regional Nature Reserve and the Sirente–Velino Regional Natural Park. When the sampling locations were overlaid on the prediction map, most fell within areas of relatively high predicted occurrence probability, an encouraging sign that the model captures real ecological signal rather than statistical artefact.

The genetic analysis, however, delivered the study’s most provocative twist. The team sequenced an 835-base-pair fragment of the mitochondrial MT-ND5 gene, a marker favoured in felid phylogenetics because it carries low homoplasy and is free of nuclear mitochondrial pseudogenes, trimming the sequences to 669 base pairs for comparison with a European reference dataset. Among 21 successfully sequenced individuals, all of which had been phenotypically classified as wildcats based on diagnostic pelage traits, the researchers found just four haplotypes grouped into three haplogroups. Only two individuals, roughly 10 percent, carried a ‘w’ haplotype reported exclusively in European wildcats. The remaining 90 percent carried maternal lineages shared with domestic cats: the dw1 haplotype accounted for 71.4 percent of samples and dw4 for 14.3 percent, while a single individual carried the domestic-type d3 haplotype.

One finding in particular stands out. The w12 haplotype, detected in two Apennine individuals, had previously been recorded in only one animal across a broad European dataset: a wildcat from Spain assigned to Felis silvestris on the basis of nuclear markers, where it was interpreted as a private Iberian lineage potentially reflecting historical isolation south of the Pyrenees. Its appearance in central Italy extends the known geographic range of this rare maternal lineage well beyond Iberia and adds to the mitochondrial diversity documented for Italian wildcats. The authors are careful, however, to note that because mitochondrial DNA is maternally inherited and represents a single non-recombining locus, this detection should not by itself be read as evidence of a direct phylogeographic connection between Italian and Iberian populations.

What, then, do the domestic-type maternal lineages mean? The answer is more nuanced than a simple hybridisation alarm. Recent genome-wide work suggests that contemporary interbreeding between wildcats and domestic cats is limited despite centuries of sympatry across Europe, which makes historical introgression a plausible contributor to the mitochondrial patterns observed here. The study’s ND5 data alone cannot establish the timing or extent of admixture; distinguishing recent hybridisation from ancient gene flow would require nuclear or genome-wide markers. There are also caveats of scale: successful sequences came from only 21 samples, and because individual identity could not be genetically confirmed for all hair samples, the apparent dominance of the dw1 haplotype may be inflated, since three hair samples sharing it were collected at the same location within a short timeframe. The haplotype frequencies should therefore be treated as preliminary.

The mismatch between pelage-based identification and mitochondrial assignment also carries a methodological warning. Diagnostic coat characteristics, such as the thick ringed tail and distinctive dorsal stripe described in classic identification keys, remain a practical field tool, but phenotypic overlap between wildcats, domestic cats, and admixed individuals means some misclassification is inevitable. That uncertainty propagates into the occurrence records underlying the distribution model itself, and the authors recommend interpreting habitat suitability patterns with this in mind. Interestingly, statistical comparison of predicted occurrence probabilities across sample categories showed that phenotypically wild individuals without genetic data occupied significantly more suitable habitat than any genetic category, hinting that the most remote, least disturbed areas may host the least admixed animals.

For conservation planners, the practical message is concrete. The Apennine chain emerges as the critical connective infrastructure for Italian wildcats, and protecting forest and grassland across suitable elevational gradients should be the priority, alongside ecological corridors and buffer zones that maintain landscape connectivity. The authors also point to precautionary measures in areas of sympatry, including responsible pet ownership and monitoring of feral cat populations to reduce wildcat–domestic cat interactions. Set against global biodiversity targets such as protecting 30 percent of terrestrial areas by 2030, the study offers a template: combine habitat modelling, updated field occurrence data, and genetic monitoring, supported by citizen science, to build an evidence base strong enough to guide action. For a species so adept at vanishing into the Mediterranean maquis, that combination of eyes and molecules may be the only reliable way to keep track of Europe’s last truly wild cat.

Subject of Research: Conservation genetics and habitat suitability modelling of the European wildcat in Italy

Article Title: Checking Wilderness Through Conservation Genetics: The Case of the European Wildcat in the Central Apennines, Italy

Article References: Maini, M., Polce, C., Bernini, F., Longeri, M., Milanesi, R., Caniglia, R., Mattucci, F., Velli, E., & Gatti, R. C. (2026). Checking Wilderness Through Conservation Genetics: The Case of the European Wildcat in the Central Apennines, Italy. Ecology and Evolution, 16(9), Article e74418. https://doi.org/10.1002/ece3.74418

Image Credits: AI Generated

DOI: 10.1002/ece3.74418

Keywords: European wildcat, Felis silvestris, conservation genetics, species distribution model, MaxEnt, mitochondrial DNA, Apennines, Italy, hybridisation, habitat suitability, camera trapping, wildlife monitoring

Finless porpoises are among the most unusual of the roughly 98 living cetacean species. Unlike most whales and dolphins, they lack a dorsal fin entirely; instead, a broad, low ridge of tubercles runs along the middle of the back, and its width has long been used to separate the wide-ridged N. phocaenoides from its narrow-ridged relatives. First described by Georges Cuvier in 1829, the genus Neophocaena was for decades treated as a single species spanning the Indo-Pacific, but molecular and morphological work eventually split it into the wide-ridged form and the narrow-ridged N. asiaeorientalis, the latter of which includes the only porpoise population in the world that lives entirely in freshwater, in China’s Yangtze River. While the narrow-ridged subspecies have attracted intense research and conservation attention, the wide-ridged species, listed as Vulnerable by the IUCN, has remained comparatively neglected, particularly the Indian Ocean populations whose genetic makeup has been almost entirely unexamined at the whole-genome level.

The Kozhikode specimen gave the team a chance to close part of that gap. Morphologically, the animal matched the diagnostic profile of N. phocaenoides: a blunt, beakless head, spade-shaped teeth, a short and broad rostrum deflected downward, and a total body length of 144 centimeters, all within the known ranges for the species. After two months of burial in salted alluvial soil, the researchers exhumed a fully articulated skeleton, now deposited as a reference specimen in the National Zoological Collections of the Zoological Survey of India’s Western Ghat Regional Centre. The skeletal preparation preserved the complete vertebral column, paired limb bones, and a skull with a condylobasal length of 185 millimeters, providing a rare anatomical voucher for a species whose remains are seldom recovered intact along the Indian coast.

The mitochondrial genome extracted from the specimen proved to be 16,386 base pairs long, with the canonical vertebrate arrangement of 37 genes: 13 protein-coding genes, two ribosomal RNA genes, 22 transfer RNA genes, and one control region. Most genes are transcribed from the heavy strand, with ND6 and eight tRNAs on the light strand, as is typical for mammals. The team documented the start and stop codons for every protein-coding gene and compared them with published mitogenomes of the narrow-ridged subspecies, finding lineage-specific variation in termination signals, including AGA and AGG stop codons in some genes. Comparative analyses across all available Neophocaena mitogenomes showed the genomes are AT-rich, that codon usage favors amino acids such as leucine, arginine, and serine, and that pairwise ratios of nonsynonymous to synonymous substitutions fall well below one for all 13 protein-coding genes, a signature of strong purifying selection that has kept these mitochondrial genes functionally stable across the genus.

The phylogenetic story proved more tangled than the tidy taxonomy might suggest. Drawing on 88 complete mitochondrial genomes from GenBank plus the new sequence, and using the harbour porpoise and Dall’s porpoise as outgroups, the researchers built a Bayesian tree that confirmed the monophyly of the porpoise family Phocoenidae with strong support. Yet the three nominal Neophocaena taxa did not sort into cleanly reciprocally monophyletic groups; instead, sequences assigned to N. phocaenoides and the two N. asiaeorientalis subspecies intermixed across the tree, with soft polytomies and extremely short branches hinting at rapid, recent diversification. Two automated species-delimitation methods, ABGD and ASAP, consistently parsed the mitogenomes into six operational taxonomic units, more than the three names currently recognized, underscoring how incomplete the genus’s classification remains.

The most striking result concerned the Arabian Sea. The newly sequenced Indian mitogenome clustered tightly with a single previously reported Arabian Sea genome, and the two together shared the highest number of segregating sites, 166, relative to their nearest neighbors of any haplotypes in the analysis. Both were assigned to the same distinct OTU by both delimitation methods. The authors suggest this reflects genuine regional genetic structuring, plausibly shaped by the Arabian Sea’s distinctive oceanography, where monsoon-driven upwelling, seasonal reversal of currents, and historically volatile glacial-age climates could have isolated porpoise populations from their Pacific counterparts. They caution, however, that sampling from the region remains thin, and that confirming the Arabian Sea porpoises as an evolutionarily significant unit will require broader population-level sampling, particularly from the Bay of Bengal, ideally using nuclear markers such as genome-wide SNPs rather than maternally inherited mitochondrial DNA alone.

To translate the genetics into conservation geography, the team built an ensemble species distribution model combining five algorithms, boosted regression trees, multivariate adaptive regression splines, generalized linear models, maximum entropy, and random forests, trained on 103 spatially rarefied occurrence records and ten uncorrelated environmental layers from the Bio-ORACLE database. The ensemble performed strongly, with training AUC values between 0.885 and 0.955 and cross-validation values between 0.862 and 0.895. One variable towered over the rest: bathymetry, which contributed nearly 62 percent of predictive power on average, reflecting the porpoise’s strict confinement to shallow coastal waters. Iron concentration, a micronutrient that fuels phytoplankton growth, contributed almost 12 percent, followed by seafloor slope, chlorophyll, and aspect. Temperature, salinity, and current velocity, by contrast, barely registered, indicating that the species’ distribution is governed primarily by depth and nutrient-driven productivity, proxies for the prey-rich coastal zones the animals depend on.

The model’s output was sobering. Within the IUCN-assessed range of roughly 2.95 million square kilometers, only about 256,742 square kilometers, or 8.71 percent, was predicted as suitable habitat, with the largest strongholds in Indonesia and China, and moderate extents in Myanmar, Malaysia, India, Vietnam, and Thailand. Worse, only 23.38 percent of the suitable area falls within existing marine protected areas or Important Marine Mammal Areas, leaving vast stretches of high-quality habitat effectively unprotected. The threat assessment layered a cumulative human impact index over the suitable habitat and found the heaviest pressures in the South China Sea region, around Singapore, China, and Taiwan, and in the Persian/Arabian Gulf, where Iraq, Kuwait, Qatar, and Iran showed substantial overlap between porpoise habitat and pollution, shipping, and coastal development. These are precisely the stressors, habitat loss to ports and aquaculture, vessel strikes in high-speed ferry corridors, and underwater noise that disrupts echolocation and foraging, that have driven the species’ decline.

The authors argue that the combination of a distinct Arabian Sea lineage and a fragmented, heavily pressured habitat portfolio demands both expanded spatial protection and smarter management. They advocate dynamic ocean management, in which protection measures shift in space and time using real-time data from remote sensing, animal tracking, and habitat models, and they call for intensified genomic sampling across the western half of the species’ range to resolve its population structure before it can inform formal taxonomic revision. The study has its limits, as the authors acknowledge: the genetic and morphological data come from a single stranded individual, and the distribution model rests on public occurrence records that capture broad patterns but not seasonal variability. Even so, the message is clear. A porpoise that slipped ashore on a Kerala beach has revealed a hidden branch of the finless porpoise family tree, and a map showing that the species’ future hangs on a sliver of shallow, productive sea that humanity has barely begun to protect.

Subject of Research: Mitogenomics and habitat suitability modeling of the Indo-Pacific finless porpoise Neophocaena phocaenoides

Article Title: Mitogenomic and Ecological Perspectives on the Wide‐Ridged Indo‐Pacific Finless Porpoise Neophocaena phocaenoides: Toward Emerging Conservation Strategies

Article References: Kamalakannan, M., Abedin, I., Putra, A., Hegde, V. D., Banerjee, D., Choi, J. H., Kim, H.-W., & Kundu, S. (2026). Mitogenomic and Ecological Perspectives on the Wide‐Ridged Indo‐Pacific Finless Porpoise Neophocaena phocaenoides : Toward Emerging Conservation Strategies. Ecology and Evolution, 16(9), Article e74249. https://doi.org/10.1002/ece3.74249

Image Credits: AI Generated

DOI: 10.1002/ece3.74249

Keywords: finless porpoise, Neophocaena phocaenoides, mitogenome, Arabian Sea, phylogenetics, species distribution model, marine conservation, habitat suitability, purifying selection, marine protected areas, cetacean genetics, IUCN Vulnerable