The work forms part of HiQBioDiv, “High-resolution quantification of biodiversity for conservation and management,” funded by Latvia’s Ministry of Smart Administration and Regional Development. The broader project is intended to develop nationwide models for European Union-protected taxa with unfavourable or unknown conservation status, as well as habitat specialists, indicator species and umbrella species. Its stated downstream applications include producing spatially explicit biodiversity layers, prioritising sites and assessing possible additions to the national protected-area network; these remain objectives of the wider project design.

The published package contains five related datasets. SpeciesList records the criteria used to consider taxa for modelling, while Nomenclature reconciles scientific names and, where available, common names across taxonomic groups and conservation lists. Observations contains the filtered occurrence records, FilteringGuide identifies land-cover classes that may indicate location errors for particular species, and EGVtable specifies candidate ecogeographical variables and their spatial scales. A shared combination of taxonomic group, assigned taxon number, scientific name and species code allows the tables to be filtered and cross-referenced.

Candidate species were drawn from nationwide taxonomic lists and from conservation-oriented groups that included protected, endangered, typical, indicator and umbrella species. Selection also depended on practical modelling constraints. Each candidate needed at least 20 expert-validated observations from 2016–2024 before spatial filtering, and taxa that could not be identified confidently from photographs were excluded. Expert judgement was used to avoid expansive, invasive, non-native, synanthropic, ecologically plastic and taxonomically difficult species whose models were considered unlikely to provide robust or conservation-relevant estimates.

Occurrence records came from the Dabasdati citizen-science portal, the governmental OZOLS nature-data system, PlutoF, the Institute of Plant Protection Research’s Agrihorts collection, several LIFE projects and personal expert databases. Taxonomic-group experts checked identifications where photographs, video or sound recordings were available and screened records for obvious coordinate errors. Records from Dabasdati or OZOLS without photographs were retained only when the observer was also an expert in the relevant taxonomic group.

Each observation retains its source identifier and a three-level relative reliability category—citizen science, expert/research or mixed. Temporal and abundance information are included when available; otherwise, a single observed individual is assumed. Coordinates are expressed in EPSG:3059, the LKS-92/Latvia TM coordinate system, and rounded to species-specific 100-, 1,000- or 5,000-metre grid-cell centres to avoid revealing sensitive locations. Only the last known observation in each grid cell is retained. Under national rules on restricted information, occurrence data for 10 taxa are omitted from the public observation file, including all bat species and the orchid Cypripedium calceolus. The rules also determine the resolutions at which observations and model projections may be published.

The FilteringGuide compares observation positions with broad Corine Land Cover classes representing built-up areas, open landscapes, tree-covered areas, wetlands and water. Coastal-dune species receive an additional check using Latvia’s exclusive economic zone and Baltic Sea coastline files. An observation in a class marked unsuitable for its species is flagged as a potential location error rather than automatically treated as invalid. The guide also records whether land-use and land-cover change analysis has been completed and which landscape classes require assessment.

For observations dated 2016–2024, the authors calculated accumulated environmental change from the year of observation through 2024. The analysis used the 30-metre Global Forest Watch v1.12 tree-cover-loss layer and annual Dynamic World land-cover classifications derived from Sentinel-2 images with cloudiness no greater than 35%. Change was aggregated at the 100-metre analysis-cell level and within landscape radii of 500, 1,250, 3,000 and 10,000 metres. Observations from 2000–2015 and 2025 were included for rare or protected species with very limited records, but these records do not contain the corresponding environmental-change calculations.

The EGV table draws on 538 candidate variables at 100-metre resolution, covering climate, hydrology, terrain, soil, Earth-observation products, landscape diversity and composition, agriculture, forestry, distance to landscape features and other land-cover classes. Species-specific selections use the 100-metre cell, an expert-estimated home-range scale and the next larger landscape scale. The framework seeks to avoid duplicating equivalent information across scales, reduce multicollinearity and spatial autocorrelation, and retain biologically interpretable gradients. For example, forest-dwelling species may use forest age and timber volume at 100 metres, broader age-class proportions at home-range scale and increasingly general forest-cover measures at landscape scale. The publication provides candidate inputs, not fitted coefficients or predicted distributions, and does not evaluate model performance, predictive accuracy or the consequences of using the data in protected-area decisions.

Subject of Research: Preparation of species occurrence data and environmental predictors for habitat suitability modelling in Latvia

Article Title: Occurrence data for habitat suitability modelling to support conservation planning in Latvia: species list, observations and ecogeographical variable selection

Article References: Starka, R., Rūsiņa, S., Auniņa, L., Balalaikins, M., Butkeviča, J., Cera, I., Čakstiņa, D., Čeirāns, A., Gailis, J., Liepiņa, L., Leimanis, I., Mežaka, A., Moisejevs, R., Ozols, J., Pilāte, D., Priedniece, I., Riters, R., Rubene, B., Strazdiņa, L., … Avotiņš, A. (2026). Occurrence data for habitat suitability modelling to support conservation planning in Latvia: species list, observations and ecogeographical variable selection. Biodiversity Data Journal, 14, Article e194577. https://doi.org/10.3897/bdj.14.e194577

Image Credits: AI Generated

DOI: 10.3897/BDJ.14.e194577

Keywords: Latvia, habitat suitability modelling, species occurrences, biodiversity data, conservation planning, citizen science, ecogeographical variables, spatial filtering, data paper, Occurrence, data, habitat

The work builds on the second edition of Chile’s vegetation classification and cartography, developed by Luebert and Pliscoff in 2017. Chilean government conservation agencies and organizations already use the national vegetation units as surrogates for ecosystems in systematic planning, including the identification of new protected areas. Researchers have also used them to examine climate-change effects, assess the impacts of Chile’s severe 2017 fires and establish categories of ecosystem risk based on International Union for Conservation of Nature guidelines. The new linkage extends that nationally useful framework toward the EcoVeg approach, which is intended to provide an international standard for classifying vegetation.

To make the correspondence, the researchers first assigned each zonal vegetation unit to an EcoVeg formation by combining coarse vegetation physiognomy with macroclimate. They then used the names and descriptions of the vegetation belts to evaluate biogeography, diagnostic species, growth forms, bioclimate, dynamics and climate-related hydrology. The main reference was the published list of American macrogroups, supplemented where necessary with proposed new units following established EcoVeg criteria. The macrogroup level was selected because the finer group level remains largely undeveloped in Latin America, although the authors believe some Chilean units may eventually fit better at that level. Chilean zonal vegetation units, or vegetation belts associated with broad climatic and geographic zones, formed the population of 125 units analyzed in the study.

Four formation classes account for Chile’s zonal vegetation. Forest and woodland vegetation occurs in central and southern Chile across climates ranging from Mediterranean to cold antiboreal conditions and includes sclerophyllous, deciduous, lauriphyllous and evergreen forests. Ten macrogroups in this formation class received 55 zonal units. Shrub and herb vegetation, discontinuously distributed through central Chile, the far north and Patagonia, includes thorny scrub, Patagonian grassland and moorland; 18 units were assigned to six macrogroups in four formations. Another 33 units fall within desert and semi-desert formations, chiefly in northern Chile but also in the Mediterranean climatic zone of the central Andes. Nineteen units of high-mountain scrub, grassland and forb vegetation were assigned to three macrogroups, bringing the zonal total to all 125 units.

The analysis proposed additions that recognize vegetation characteristics not adequately represented in the existing international list. The authors introduced a Mediterranean Forest and Woodland formation and a Chilean Mediterranean Forest and Woodland division to accommodate the sclerophyllous and deciduous forests of central Chile. They also proposed the Valdivian Temperate Conifer Forest and Woodland macrogroup within the Cool Temperate Forest and Woodland formation. Two additional units, Magellanian Antiboreal Deciduous Forest and Magellanian Antiboreal Evergreen Forest, were needed because the existing classification did not distinguish both evergreen and deciduous Magellanian forests under an antiboreal bioclimate. These proposals extend the international framework rather than merely applying it unchanged, providing specific units for vegetation that the authors concluded was either absent or insufficiently differentiated in the earlier macrogroup list.

Several mismatches emerged where physiognomy, climate and biogeography do not align neatly. Chilean formations and International Vegetation Classification formations are both largely based on physiognomic and ecological criteria, but the international system combines multiple previous classification approaches. The authors found a particular difficulty in deciding whether macroclimate or biogeography should take precedence at different levels. For example, forests dominated by Nothofagus pumilio and Nothofagus betuloides can display similar structure and occur in both temperate and antiboreal climates. They are placed in different regional divisions because of biogeography, then distinguished as deciduous and evergreen macrogroups through growth form and dominant species. The authors also questioned the global suitability of terms such as “Boreal” and “Alpine” in formations that occur in the Southern Hemisphere, suggesting that such names and others with the same problem should be revised to make the international system more globally applicable.

Central Chile provided a further test of the hierarchy because its Mediterranean-climate forests were accommodated within a broader warm-temperate formation that combined sclerophyllous and deciduous vegetation. The proposed Chilean division reflects differences in floristic composition and phylogenetic diversity from other parts of the country, and the authors suggest it could eventually include appropriate Mediterranean-climate units elsewhere, including parts of California, the Mediterranean basin, South Africa and Australia. They also proposed placing Mediterranean high-Andean scrub and grassland in the cool semi-desert formation rather than the tropical high-mountain formation used in the existing list. These judgments expose differences among climate definitions, including the seasonal Mediterranean bioclimate used in the Chilean classification and systems in which Mediterranean climates are classified within temperate zones. The authors note that alternative definitions may differ in hierarchical position and spatial extent, so these assignments are not entirely independent of the climatic system adopted.

Not every existing macrogroup could be matched cleanly. Some were absent from the Chilean inventory, including Southern Andean Montane Salt Marsh and Sechura Atacama Semi-Desert Cliff and Pavement, while others could not be identified confidently from the national descriptions. Xeric Puna succulent scrub, represented in Chile by locally abundant Trichocereus atacamensis vegetation, had been embedded within broader high-elevation units rather than recognized separately. The division of Chile’s coast into Atacama and Peruvian Sechura components remains potentially contentious because the deserts have been described as floristically very different in some studies, although the authors cite phylogenetic and floristic evidence for broader relationships and retained the wider biogeographical treatment. These examples show why international classification requires explicit decisions about boundaries and hierarchy, not just a direct one-to-one translation of national maps.

Vegetation that does not follow the climatic zonation of the landscape proved more difficult. The researchers provisionally assigned intrazonal units to 23 macrogroups, 23 divisions and 17 formations, but some correspondences are doubtful because Chile’s intrazonal categories are broad and lack sufficiently detailed descriptions. Aquatic vegetation, including floating, natant, submerged, semi-aquatic and ephemeral-wetland communities, and open-rock vegetation on coasts, cliffs, inland outcrops, landslides and caves were also considered. Moorlands in southern Chile could not be placed in a dedicated macrogroup and were associated with bog and fen categories, even though they may be more closely related to azonal wetlands. Units dominated by Nothofagus antarctica also resisted formation-level placement because the species may grow as a large shrub or small tree yet strongly determines the vegetation’s physiognomy. The researchers included these units in forest formations because accommodating them in a new macrogroup could require major changes to the international classification.

Overall, the study shows that most of Chile’s mapped zonal vegetation can be represented within the EcoVeg hierarchy, while the proposed new formations, divisions and macrogroups identify places where the international system needs adjustment. The accompanying macrogroup map is available as a shapefile, and the authors present it as a tool that can facilitate collaboration across borders. A separate effort is reportedly under review to translate Chilean phytosociological units directly into International Vegetation Classification units down to association level, potentially adding secondary, degraded and ruderal vegetation. The present results remain a cartographic and classificatory contribution: they do not demonstrate improved conservation outcomes, and the authors explicitly characterize the intrazonal assignments as preliminary. Describing and classifying those non-zonal communities remains a major pending task in Chilean vegetation science.

Subject of Research: International classification and mapping of Chilean zonal and intrazonal vegetation

Article Title: The vegetation of Chile and the EcoVeg approach in the context of the International Vegetation Classification project

Article References: Luebert, F., & Pliscoff, P. (2022). The vegetation of Chile and the EcoVeg approach in the context of the International Vegetation Classification project. Vegetation Classification and Survey, 3, 15-28. https://doi.org/10.3897/vcs.67893

Image Credits: AI Generated

DOI: 10.3897/VCS.67893

Keywords: Chilean vegetation, EcoVeg, International Vegetation Classification, vegetation mapping, biogeography, ecosystem classification, Mediterranean forests, Patagonian grasslands, Magellanian forests, conservation planning, vegetation, Chile

The study, led by researchers examining 13 marine taxonomic groups, draws on more than 14 million occurrence records from the Ocean Biogeographic Information System. Rather than treating biodiversity maps as fixed products, the team reconstructed how these maps evolved over time by partitioning records chronologically and comparing patterns generated from progressively larger subsets of data. This approach allowed the researchers to quantify the degree of agreement between early, data-poor maps and the most recent, data-rich versions, producing a convergence score for each taxonomic group and each biodiversity metric.

The framework focuses on three distinct biodiversity patterns: species richness, latitudinal diversity gradients, and rarity-weighted species richness, a metric that emphasizes species with restricted ranges. Species richness simply counts the number of species present in each grid cell, while latitudinal diversity gradients capture the well-known tendency for species numbers to peak in tropical waters. Rarity-weighted species richness, by contrast, highlights regions harboring species with limited distributions, making it particularly relevant for conservation prioritization.

To generate species distributions from occurrence records, the researchers combined convex hull polygons with environmental filtering based on sea surface temperature and depth. Records deviating by more than three standard deviations from the range of environmental conditions occupied by a species were treated as outliers and removed. The team also accounted for geographic sampling bias by partitioning data in a way that balanced increases in spatial coverage, rather than simply drawing random subsets that could overstate convergence.

The results show that convergence increased with the volume of available data and with the spatial and taxonomic breadth of sampling. Groups such as Actinopterygii, Chondrichthyes, and Scleractinia, which benefit from extensive records and broad sampling, achieved convergence scores of 0.66 or higher for species richness and latitudinal diversity patterns. These findings suggest that the biodiversity maps for these groups are approaching a stable state where additional data are unlikely to substantially alter the overall picture.

In contrast, groups with fewer records or highly restricted sampling showed markedly lower convergence. Scaphopoda and Polyplacophora, two of the most data-poor groups examined, scored 0.29 and 0.51 respectively for species richness, indicating that their current maps remain highly sensitive to new discoveries. Marine mammals and marine reptiles, despite having relatively few species, also exhibited weak convergence, reflecting the disproportionate influence that each individual species has on overall patterns when species numbers are small.

Rarity-weighted species richness showed the weakest convergence across all groups, with scores ranging from 0.19 to 0.51. This metric proved especially sensitive to the addition of new records, because restricted-range species are disproportionately likely to be discovered in previously unsampled areas. The findings suggest that conservation maps based on rarity remain highly provisional across all marine taxonomic groups examined, underscoring the need for continued sampling in underexplored regions.

The researchers also found that convergence was higher when mapped species richness showed little correlation with the density of sampling effort. When species richness closely tracks sampling intensity, patterns are more likely to reflect where people have looked rather than where species actually live. Conversely, when richness decouples from sampling density, it suggests that patterns are approaching a state where they reflect genuine biogeographical structure rather than artifacts of uneven exploration.

The study’s authors note that their framework assumes convergence, meaning that patterns approach a stable state as data accumulate. They acknowledge that this assumption may not hold for all groups, particularly those undergoing rapid environmental change or experiencing significant taxonomic revision. Nevertheless, they argue that the approach provides a useful diagnostic for identifying which biodiversity maps are ready for use in conservation and macroecological applications and which require further data collection before they can be considered reliable.

The findings carry practical implications for conservation planning, biogeographical research, and the design of future biodiversity surveys. By quantifying the maturity of biodiversity knowledge across taxonomic groups, the framework can help prioritize sampling efforts in regions and for groups where additional data are most likely to improve the reliability of existing maps. The researchers suggest that their approach could be extended to terrestrial systems and to other biodiversity metrics, offering a general tool for assessing the readiness of biodiversity data to support evidence-based decision-making.

Subject of Research: Quantitative assessment of the reliability and convergence of marine biodiversity maps across taxonomic groups

Article Title: The maturity of marine biodiversity maps: insights from the convergence of global patterns

Article References: Ikari, S., Shiono, T., & Kubota, Y. (2026). The maturity of marine biodiversity maps: insights from the convergence of global patterns. Frontiers of Biogeography, 19, Article e178549. https://doi.org/10.21425/fob.19.178549

Image Credits: AI Generated

DOI: 10.21425/fob.19.178549

Keywords: marine biodiversity, species richness, latitudinal diversity gradient, rarity-weighted species richness, convergence analysis, conservation planning, biogeography, Ocean Biogeographic Information System, taxonomic groups, sampling bias, data-poor taxa, biodiversity mapping

The research team sampled 16 sites distributed across the basin’s main watercourses, from near sea level up to 1,062 metres in the Rif Mountains. The sites were grouped into four categories reflecting altitude and land use: natural and semi-natural upstream reaches, semi-natural and agricultural midstream sites, agricultural lower reaches, and heavily urbanised downstream stations. Sampling covered four campaigns between April 2017 and February 2018, spanning spring, summer, autumn, and winter, and thus capturing both the wet Mediterranean season, when rain falls mainly between October and April, and the long dry period from May to September.

In total, the surveys yielded 1,268 individual beetles representing 50 species from 10 families and 27 genera. Diversity was strikingly uneven. The richest site, Tkaraa, lies inside Bouhachem Natural Park and hosted 22 species, while the most abundant site, Zarka, produced 194 individuals. At the opposite extreme, urbanised downstream sites such as Taboula and Roumana, close to Tétouan, yielded only a single species each. Gyrinidae, the whirligig beetles that skim the water surface in swirling groups, dominated numerically with 54 percent of all individuals, yet the diving beetles of the family Dytiscidae proved the most diverse, contributing 17 species, with the genus Agabus the richest in the basin.

Seasonality left a clear signature on the assemblages. Species richness peaked in spring 2017 with 31 species and 377 individuals, whereas the post-flood winter of 2018 produced only 208 individuals from 19 species. The two great beetle suborders responded in opposite ways. Adephaga, which include the whirligigs and diving beetles, dominated the dry period, averaging 20.3 occurrences per station in spring and favouring warm, standing pools. Polyphaga, including the riffle beetles of Elmidae and the minute moss beetles of Hydraenidae, were more frequent under wetter, flowing conditions, reaching an average of 9.38 occurrences per station in autumn. A factorial correspondence analysis separated the fauna cleanly by season, with rheophilic species tied to wet-season flows, thermophilic species concentrated in summer pools, and eurythermic generalists associated with spring’s relatively stable conditions.

Functional traits reinforced the pattern. Predators made up 73 percent of individuals and swimmers 72 percent, although the authors caution that this signal is heavily skewed by the sheer abundance of surface-dwelling whirligig beetles, which count as both. When Gyrinidae were excluded, the picture changed: shredders preferred winter, clingers replaced swimmers as the dominant group in wetter months, and the assemblage shifted toward benthic, substrate-associated forms. The physiological explanation is elegant. Many Adephaga breathe atmospheric air stored beneath their elytra, allowing them to endure warm, oxygen-poor isolated pools as drought advances. Riffle beetles and their relatives, by contrast, depend on plastron or cuticular respiration and require well-oxygenated running water, so they suffer when flow ceases and dissolved oxygen collapses.

The environmental gradient along the river was stark. Mean water temperature rose from 15.5°C at upstream sites to 20.8°C downstream, and electrical conductivity climbed from 269 to 936 microsiemens per centimetre. Dissolved oxygen fell from 7.84 to 2.49 milligrams per litre, while biochemical oxygen demand increased from 10.7 to 32.3 milligrams per litre and chemical oxygen demand from 24.8 to 74.6. Beetle abundance mirrored this degradation, plunging from a mean of 29.6 individuals per upstream site to just 1.19 at the most impacted stations, with species richness dropping from 5.5 taxa to 0.81. A co-inertia analysis, statistically confirmed by Monte Carlo permutation tests, captured this structure: the first two axes explained over 94 percent of the joint variation and opposed well-oxygenated upland refuges dominated by sensitive rheophilic species against lowland, organically enriched reaches occupied by tolerant generalists such as Laccophilus minutus and Helophorus atlantis.

Perhaps the most consequential finding concerns where the compositional break occurs. A permutational multivariate analysis of variance showed that habitat type explained 33.7 percent of the variation in community composition, but pairwise comparisons revealed that urban sites differed significantly from every other category, while upstream, midstream, and agricultural sites did not differ significantly from one another. In other words, moderate agricultural pressure leaves beetle assemblages broadly continuous along the river, but urbanisation produces a threshold-like rupture. Beta-diversity partitioning added a second insight: total dissimilarity across sites was high, at 0.701, and species turnover accounted for 73.7 percent of it, meaning that communities along the Martil are structured mainly by the replacement of specialist species by generalists, not simply by the progressive loss of species.

The study also carries a warning for endemic species. Helophorus atlantis, a Moroccan endemic classified as highly vulnerable, was recorded only in the polluted downstream zone. The authors note that in intermittent rivers, floods can act as mass dispersal events that sweep upstream individuals into lowland reaches, so such records may reflect transient, trapped populations rather than genuine ecological tolerance. Whether this beetle truly survives under urban pollution or was merely drifting through remains an open question, and one that matters for assessing how North Africa’s endemic aquatic insects cope with disturbance. Meanwhile, sites within Bouhachem Natural Park and permanent midstream watercourses showed high endemism, particularly among Polyphaga, confirming their role as critical refuges.

These results resonate across the wider Mediterranean, a global freshwater biodiversity hotspot where climate-driven aridification is lengthening dry phases and human pressure is fragmenting the remaining permanent waters. As riffles disappear and pools isolate, desiccation-tolerant, air-breathing, actively dispersing species gain ground while oxygen-dependent specialists retreat upstream, a dynamic that risks biotic homogenisation outside protected areas. In the Rif Mountains, one of Morocco’s principal aquatic endemism hotspots, the stakes are especially high, yet national knowledge of aquatic beetle distributions and the ecological requirements of endemic taxa remains incomplete, and IUCN criteria designed primarily for vertebrates often fail to capture the needs of invertebrates.

The authors’ central recommendation is practical: conservation planning should extend beyond headwater parks to include the permanent mid-reaches where sensitive and endemic species persist, and pollution and hydromorphological degradation in lowland reaches must be reduced to restore longitudinal connectivity. Because water beetles respond so sharply to flow permanence, oxygen levels, and organic pollution, they can serve as effective bioindicators for identifying priority sites and guiding the expansion of protected area networks through complementarity-based approaches. With more than half of the Martil basin already protected but its lowland waters still deteriorating, the message from these 1,268 beetles is unambiguous: the river’s future diversity will be decided not only in its mountain strongholds, but in the connecting reaches that urban Morocco has so far allowed to degrade.

Subject of Research: Effects of seasonal river drying and human pollution on aquatic beetle diversity in a Mediterranean river basin in northern Morocco

Article Title: Effects of hydrological intermittency and anthropogenic pressure on aquatic beetle assemblages in the Martil River (north-western Morocco)

Article References: Guellaf, A., Bennas, N., & Kettani, K. (2026). Effects of hydrological intermittency and anthropogenic pressure on aquatic beetle assemblages in the Martil River (north-western Morocco). Biodiversity Data Journal, 14, Article e192023. https://doi.org/10.3897/bdj.14.e192023

Image Credits: AI Generated

DOI: 10.3897/BDJ.14.e192023

Keywords: aquatic beetles, hydrological intermittency, Mediterranean freshwater, Morocco, Rif Mountains, urban pollution, biodiversity indicators, species turnover, water quality, conservation planning, endemism, macroinvertebrates

The tool in question is the mobility oriented parity metric, or MOP, first introduced to quantify how dissimilar the environments of a transfer area are from those of the area where a model was calibrated. The metric works by measuring multivariate environmental distances between the conditions of interest and a predefined percentage of the closest reference conditions in environmental space. Unlike approaches that compare new conditions only to the average or centroid of the calibration environment, MOP accounts for the density of reference conditions, making it uniquely sensitive to how environmental combinations are actually distributed across the calibration area. Where conditions fall outside the ranges seen during calibration, extrapolation may produce misleading results, especially if the model’s response curves were truncated at the edges of the calibration data.

Despite its popularity, MOP has long suffered from a frustrating limitation. When it flags a location as environmentally non-analogous, it says nothing about which variables are responsible, whether their values are too high or too low, or which combinations of variables create the novelty. Earlier tools such as the multivariate environmental similarity surface, proposed by Elith and colleagues, could identify the single most extreme variable, while another approach known as ExDet could point to the most novel variable or variable combination, but neither offered a complete picture. No existing method simultaneously identified every variable outside its reference range, distinguished whether extreme values sat at the high or low end, and revealed which variable combinations produced novel conditions. The new study, led by researchers at the University of Kansas Biodiversity Institute, closes precisely that gap.

To demonstrate the extended metric, the team ran three computational challenges using two geographic scenarios. The first compared the environmental conditions of the Lower 48 United States, serving as the calibration area, with a vast transfer region spanning most of the Americas. The second mimicked a realistic modeling exercise for the Transvolcanic Jay, a bird whose accessible calibration area was estimated through dispersal simulations, projected onto the whole of Mexico. Environmental data came from six bioclimatic variables at ten arc-minute resolution, including maximum temperature of the warmest month, minimum temperature of the coldest month, temperature annual range, precipitation of the wettest and driest months, and precipitation seasonality. These variables were chosen because they capture environmental extremes and seasonality, properties that are biologically meaningful for a wide range of organisms.

The first challenge produced strikingly detailed maps of novelty. In the Americas example, non-analogous conditions toward the low end of temperature variables appeared at high latitudes, reflecting the frigid winters of Canada and Patagonia that have no counterpart in most of the calibration region. Novelty toward the low end of temperature annual range emerged near the Equator, where thermal regimes are far more stable than in the temperate calibration area, while the hyper-arid Atacama Desert stood out for precipitation variables. Novel conditions toward the upper end of most variables clustered at low latitudes in the tropics, with extensions into dry regions for precipitation seasonality and into zones with extreme winters for temperature annual range. In the jay example, novelty toward the low end appeared mainly in northern and northeastern Mexico, driven by precipitation of the wettest month, while high-end novelty tracked temperature annual range in the country’s north.

The practical implications of this granularity are considerable. Knowing that a particular temperature variable shows large areas of non-analogous conditions in a future climate scenario gives modelers an immediate warning of extrapolation risk. If two candidate predictors carry similar information in the model, MOP results can tip the balance toward the one that stays within calibration bounds. Even more subtly, the interaction between novelty direction and model response curves matters: if a model’s predicted suitability decreases toward the high end of a novel variable, spurious extrapolation is unlikely, but if suitability increases in that direction, the risk rises sharply. This makes reading MOP outputs alongside the model’s response curves essential for robust assessments of prediction reliability, a point the authors emphasize throughout their discussion.

The second challenge probed how the density and position of reference conditions in environmental space shape MOP results. The team defined six rectangular blocks of equal area within a two-dimensional environmental space defined by minimum temperature and precipitation: two inside the reference cloud, close to high-density and low-density regions respectively; two outside the cloud but within reference ranges; one just beyond the range of a single variable; and one far outside the cloud in both variables. The results confirmed that point density exerts a direct effect on dissimilarity values. Distances measured inside the reference cloud were smaller where density was high, and conditions located outside the cloud but within reference ranges sometimes produced distances comparable to, or larger than, conditions outside the range of one variable, simply because reference points were sparser nearby.

Parameter choices also proved consequential. The researchers tested Euclidean and Mahalanobis distances, with and without centering and scaling of variables, and with reference percentages of 10, 5, and 1 percent of the calibration cloud. Mahalanobis distances, which incorporate a variance-covariance matrix derived from all reference conditions, better distinguished broadly far conditions from near ones, while Euclidean distances offered finer differentiation among regions close to the reference cloud. Lower reference percentages reduced dissimilarity values within reference ranges, and rescaling distances to a zero-to-one range made comparisons across treatments clearer. Based on these patterns, the authors recommend using a relatively large percentage of reference conditions, around ten percent, together with scaled variables and Euclidean distances when the goal is a clear separation of conditions outside the reference cloud.

The third challenge examined a factor that had never been tested for MOP before: the size of the sample used to represent the calibration area. Comparing the full set of more than 40,000 calibration cells against subsets of 20,000, 10,000, and 5,000, the team found that sample size had its most dramatic effect on the identification of non-analogous areas rather than on overall dissimilarity patterns. With only 5,000 points, roughly a tenth of the full data, the number of variables detected as non-analogous increased noticeably in the Amazon region, particularly toward the low end of temperature annual range. Rare environmental combinations are more likely to be captured by larger samples, so undersampling the calibration area can mask genuine extrapolation risks. Computation times fell steeply with smaller samples, from about 141 seconds for the full dataset to under 20 seconds for 5,000 points, but the authors argue that computing time should not drive the choice of sample size given steadily improving technology.

The study arrives at a moment when ecological niche models are being applied to ever more urgent questions, from anticipating range shifts under climate change to predicting where invasive species and disease vectors might establish next. In all of these applications, the difference between an informed projection and a spurious extrapolation can determine whether scarce conservation resources are allocated wisely or squandered. The extended MOP metric, implemented in a new R package available on CRAN with full code hosted on GitHub, offers researchers a quantitatively efficient means of characterizing extrapolation uncertainty with unprecedented detail. By revealing which variables drive novelty, in which direction, and in which combinations, it transforms a blunt alarm bell into a diagnostic instrument. The authors’ broader message is clear: careful, detailed comparison of calibration and transfer environments is not an optional refinement but a best practice, one that should shape variable selection, background sampling, and the interpretation of every model transferred beyond the conditions that trained it.

Subject of Research: Improved quantification of environmental dissimilarity and extrapolation risk in ecological niche model transfers

Article Title: Detailed multivariate comparisons of environments with mobility oriented parity

Article References: Cobos, M. E., Owens, H. L., Soberón, J., & Peterson, A. T. (2024). Detailed multivariate comparisons of environments with mobility oriented parity. Frontiers of Biogeography, 17, Article e132916. https://doi.org/10.21425/fob.17.132916

Image Credits: AI Generated

DOI: 10.21425/fob.17.132916

Keywords: ecological niche models, mobility oriented parity, model transfer, extrapolation risk, biogeography, climate data, environmental dissimilarity, species distribution modeling, conservation planning, bioclimatic variables, Mahalanobis distance, sample size

Roads are among the most pervasive drivers of habitat fragmentation on Earth, and amphibians suffer disproportionately. Because these animals depend on both aquatic and terrestrial habitats, move slowly, and often undertake explosive breeding migrations that force mass road crossings, a single stretch of pavement can slice through the entire annual cycle of a population. Conservationists have long known that proximity to water predicts amphibian roadkill, but most studies stop there, relying on coarse, generalized predictors applied to all species at once. The new research, led by a team analyzing five understudied frog species in one of Taiwan’s richest amphibian hotspots, shows just how much that approach misses.

The study combined nearly two years of standardized roadkill surveys — 33 motorbike and walking surveys conducted at speeds under 10 kilometres per hour — with citizen science records from Taiwan’s Roadkill Observation Network, yielding 234 amphibian roadkill records along a road that climbs from 240 to 360 metres in elevation, flanked by steep forested hills on one side and the Ruizi Creek on the other. Five target species were selected for modeling, including the vulnerable Rana sauteri, whose mass mortality events were documented by volunteer observers, and Buergeria robusta, a stream-breeding frog whose May and June breeding migrations are well known to local frog-rescue volunteers.

The methodological innovation lies in the analysis. Rather than pooling all species together, the team built ten separate models: one multispecies model, five single-species models, and four temporal models for Buergeria robusta split across its breeding season. Environmental predictors were measured within buffer zones radiating from 25 to 250 metres around each roadkill point, and — crucially — the buffers were split into uphill and downhill halves, because a road running along the base of a steep valley presents fundamentally different environments on each side. The predictors ranged from stream order and forest cover to fine-scale features rarely examined before: the number of streetlights, the length of Jersey barriers and W-beam guardrails, the condition of open drainage channels, and even the presence of bus stops and leaf litter.

To ensure the results were robust, the researchers used a dual-model consensus approach, requiring that predictors be retained by both a Random Forest model, which captures non-linear interactions, and a Logistic Regression model, which yields interpretable linear relationships. Class imbalance was handled with synthetic over-sampling and class weighting, and predictor preselection using harmonic mean p-values, correlation filtering and variance inflation factors trimmed the variable set before modeling. Only predictors surviving both algorithms were accepted as consensus findings — 24 in total, evenly split between habitat-related and road-related factors.

The performance gap between model types was striking. The multispecies model achieved test AUC values of only 0.552 to 0.610, barely better than chance, while single-species models reached 0.746 to 0.827 and the temporal models of Buergeria robusta peaked at 0.851. In other words, lumping species together destroys the very signal that matters. Each species responded to a distinct set of environmental drivers, and even within a single species the relevant predictors and their spatial scales shifted dramatically across the breeding season. The Hylarana latouchii model, notably, produced no consensus predictors at all, likely reflecting that species’ sedentary, generalist ecology and small sample size.

Water features emerged as the most consistent drivers, but with a directional twist. Downhill rivers and larger stream networks channel amphibian movements toward the road and raised mortality risk, while high-quality uphill streams were associated with fewer crossings — suggesting that when suitable habitat lies upslope, animals simply do not need to reach the pavement. For Odorrana swinhoana, a forest-dwelling species, uphill broadleaf forest increased risk, while for Buergeria robusta, downhill cropland proved deadly during the breeding season. Streetlights told an even more nuanced story: light density was positively associated with roadkill for Rana sauteri and for Buergeria robusta in June, yet negatively associated with roadkill for the same frog species in May, possibly because brighter ambient light in downhill foraging zones reduced the need to approach the road at all. This is the first documentation of artificial lighting shaping the spatial concentration of mass-migration crossings in these species.

Structural features of the road itself proved equally consequential, and their effects could flip depending on design and season. Long Jersey barriers and tall uphill retaining walls reduced roadkill by blocking access to the road, echoing findings for common toads in Portugal. But long downhill retaining walls and W-beam guardrails increased mortality, apparently by trapping migrating frogs or funneling them through gaps into traffic. Drainage channels played dual roles as well: vegetated channels acted as ecological traps that attracted animals, while longer waterlogged channels served as barriers that prevented crossings entirely. These details matter because they are exactly the features engineers control when building or retrofitting a road.

What elevates the study beyond a regional case study is its translation into practice. The authors present a hierarchical mitigation framework spanning the full life cycle of infrastructure projects. At the planning stage, road alignments should avoid critical habitats, maintaining buffers of at least 50 metres from large downhill streams and 25 metres from uphill forests, with underpasses and drift fences prioritized near stream crossings. At the design stage, streetlights can be fitted with amber or red spectra, shielding, and seasonal dimming schedules aligned with migration peaks; retaining walls can be retrofitted with escape ramps and permeable sections; drainage channels can be engineered with sloped sides so trapped animals can climb out. During predictable mass migrations, dynamic signage, temporary speed reductions, short-term road closures and volunteer-run bucket-trap fencing offer immediate relief, while ongoing maintenance — clearing drains, inspecting underpasses, adjusting lighting schedules — keeps permanent measures functional.

The framework’s real promise is transferability. Because it is organized around mechanisms — directional hydrology, structural permeability, landscape composition, illumination effects — rather than species labels, the authors argue it can be adapted to roads, railways, canals and energy corridors from arid to subtropical environments, and even extended to pedestrian safety mapping. The study’s limitations are acknowledged: the datasets span less than two years, dynamic variables such as rainfall and traffic volume were not included, and interannual variation remains unmeasured. Yet the core message is unambiguous and, for conservation planners, quietly revolutionary. One-size-fits-all hotspot maps and generalized mitigation guidelines are insufficient. Effective protection of the world’s most imperiled vertebrates demands models tuned to individual species, individual seasons, and individual roadside structures — down to the streetlight, the wall, and the ditch.

Subject of Research: Environmental and seasonal drivers of amphibian roadkill on a Taiwanese mountain road

Article Title: Precise roadkill environmental factors identification for sustainable planning and design through spatial and temporal modelling

Article References: Karanasios, P. I., Schmeller, D. S., Hsu, Z.-T., & Lin, Y.-P. (2026). Precise roadkill environmental factors identification for sustainable planning and design through spatial and temporal modelling. Nature Conservation, 61, 1-27. https://doi.org/10.3897/natureconservation.61.171624

Image Credits: AI Generated

DOI: 10.3897/natureconservation.61.171624

Keywords: amphibian roadkill, habitat fragmentation, Random Forest, logistic regression, streetlights, drainage channels, breeding migration, Taiwan, multiscale modeling, road ecology, conservation planning, citizen science

The Greater Kafue Ecosystem comprises Kafue National Park and nine surrounding Game Management Areas, covering approximately 66,000 square kilometers. Between 1984 and 2024, the proportion of natural habitat in the region declined from approximately 98.31% to 83.96%, a transformation that has been spatially heterogeneous across the landscape. Some Game Management Areas have retained over 96% of their natural cover, while others have fallen below 40%, indicating that local factors interact with regional pressures to produce distinct outcomes. This variability highlights the need for targeted interventions rather than uniform conservation policies across the entire ecosystem.

To address this complexity, the researchers employed a Random Forest classifier, an ensemble learning algorithm known for its ability to handle high-dimensional data and capture non-linear relationships. The model was trained on a binary response variable derived from Landsat satellite classifications for 1984 and 2024, which distinguished between natural cover types such as forest and grassland, and anthropogenic types such as cropland and built-up areas. Fourteen predictor variables were evaluated, including population density, distance to settlements, mean annual rainfall, elevation, and the proportion of cultivated land within a one-kilometer neighborhood.

The model demonstrated strong predictive performance on an independent, spatially blocked test set, achieving an accuracy of 0.77 and an Area Under the Curve (AUC-ROC) of 0.83. The F1-score of 0.81 and Cohen’s Kappa of 0.53 further confirmed the model’s reliability, with ten-fold spatial cross-validation showing stable results. A higher false-positive rate compared to the false-negative rate indicated a precautionary bias, which is particularly suitable for conservation screening where identifying potential risk areas is prioritized over missing them.

Among the fourteen predictors, the percentage of cultivated land emerged as the dominant factor, accounting for 33.6% of the total variable importance. This was followed by mean annual rainfall at 12.0% and distance to settlements at 11.89%. The analysis revealed that socio-economic factors collectively contributed approximately 47.3% of the predictive power, significantly outweighing climatic, topographic, and conservation status variables. This finding aligns with broader evidence that smallholder agricultural expansion is the primary pathway of habitat conversion in sub-Saharan African woodlands.

The study identified specific cut-points where the predicted probability of habitat loss changes sharply. For cultivated land, a threshold of approximately 10% cover within a local area marked a significant increase in susceptibility, with loss probability rising from 11.3% below this level to 79.9% above it. Similarly, areas within 4 kilometers of settlements showed a high-risk influence zone, with a 75.1% probability of loss compared to 25.7% beyond that distance. These statistical cut-points serve as operational early-warning indicators for conservation managers, although they are not validated ecological thresholds.

Spatial susceptibility mapping revealed pronounced contrasts across the ecosystem. The Game Management Areas of Bilili, Namwala, and Mumbwa were identified as the principal high-susceptibility zones, with Bilili showing the highest risk at 94.5% of its area classified as high or very high susceptibility. In contrast, Kafue National Park remained predominantly low risk, with 95.4% of its area classified as very low susceptibility. This spatial differentiation supports a strategy of targeted monitoring and preventive action in the most vulnerable Game Management Areas, while recognizing that the national park remains relatively stable.

The researchers also conducted a sensitivity analysis to address potential endogeneity in the cultivated-land predictor, which is temporally close to the 2024 outcome. When the model was refitted without this variable, performance declined but remained well above chance, indicating that the predictive skill is not solely dependent on this single factor. However, the dominance of socio-economic predictors was conditional on the inclusion of cultivated land, with climatic and proximity variables becoming more prominent in its absence. This nuance is crucial for interpreting the results and designing robust conservation strategies.

The study acknowledges several limitations, including the cross-sectional nature of the analysis, which prevents the establishment of temporal causation. Most socio-economic predictors represent conditions around 2020–2021, while the response integrates conversion over four decades. Additionally, the model’s reliance on a companion Landsat classification means that thematic and change-detection errors propagate into the results. Despite these constraints, the study provides a valuable methodological template for applying machine learning to conservation challenges in other African multi-use landscapes.

By integrating the Human-Environment Interaction and Driver-Pressure-State-Impact-Response frameworks with interpretable machine learning, the study offers a comprehensive view of habitat loss drivers in the Greater Kafue Ecosystem. The findings underscore the critical role of agricultural expansion and accessibility in shaping conservation outcomes and provide actionable metrics for prioritizing interventions. As the ecosystem continues to face pressures from population growth and land-use change, this spatially explicit approach can help guide effective and targeted conservation efforts to protect its rich biodiversity.

Subject of Research: Drivers of habitat loss in Zambia's Greater Kafue Ecosystem

Article Title: Drivers of forest loss and habitat degradation in Zambia’s Greater Kafue Ecosystem

Article References: Mulenga, G., Phiri, D., Simwanda, M., & Nyirenda, V. R. (2026). Drivers of forest loss and habitat degradation in Zambia’s Greater Kafue Ecosystem. Discover Forests, 2(1), Article 81. https://doi.org/10.1007/s44415-026-00143-3

Image Credits: AI Generated

DOI: 10.1007/s44415-026-00143-3

Keywords: habitat loss, Random Forest, Zambia, Greater Kafue Ecosystem, agricultural expansion, conservation planning, remote sensing, Miombo woodlands, spatial analysis, biodiversity, Drivers, forest

The central insight behind RIRC is deceptively simple: knowing how damaging a species can be is not the same as knowing where it will actually do damage. Invasion risk, the researchers argue, is not an inherent property of a species but an emergent outcome of interactions between invader traits and the ecological and socioenvironmental characteristics of the receiving region. The same non-native plant may cause severe ecological disruption in one province while remaining rare and inconsequential in another, because impacts depend on biotic interactions, abiotic conditions, and human pressures that all vary across the landscape. Trait-based or globally derived assessments alone, therefore, cannot reliably predict local ecological outcomes.

To bridge this gap, the framework combines two established tools. The first is the Environmental Impact Classification for Alien Taxa, known as EICAT, a standardized system developed under the International Union for Conservation of Nature that ranks alien species from Minimal Concern to Massive impact based on documented harm to native biodiversity. The second is species distribution modeling, or SDM, which uses environmental variables such as climate, topography, and land cover to predict where a species could establish and spread. By multiplying the two scores on aligned one-to-five scales, RIRC produces a regional risk value ranging from one to twenty-five, which is then binned into six categories from Very Low to Extreme. The multiplicative approach is deliberately conservative: an area scores high risk only when both impact severity and climatic suitability are high, avoiding the compensatory behavior of additive methods where a high value in one dimension can mask a low value in the other.

The technical machinery behind the maps is substantial. The researchers compiled verified occurrence records from the Global Biodiversity Information Facility, the CABI Invasive Species Compendium, and Iranian literature in both English and Persian, applying rigorous quality control that retained only georeferenced records with spatial uncertainty of a kilometer or less and thinned the data to a single record per square kilometer. Climatic suitability was modeled using random forest algorithms trained on nineteen bioclimatic variables from the CHELSA database at roughly one-kilometer resolution, with multicollinearity filtered out through Spearman rank correlations and variance inflation factor analysis. Pseudo-absence points were generated at a one-to-one ratio to presence records and placed at least two degrees away from known occurrences, a strategy shown to optimize random forest performance for species with limited data. Hyperparameters were tuned through five-fold cross-validation with spatially structured folds, and model performance was assessed using out-of-bag error, area under the receiver operating characteristic curve, and the True Skill Statistic.

The models performed remarkably well. Test accuracies ranged from 90.3 percent for snake vine to 99.2 percent for golden wreath wattle, with twelve species exceeding 95 percent. AUC values varied between 0.928 and 0.992, and TSS values ranged from 0.857 to 0.984. Notably, predictive performance did not differ systematically by growth form, photosynthetic pathway, habitat preference, or EICAT impact magnitude, suggesting the framework is robust across ecological types. Predicted high-suitability areas varied enormously among species, from 18,432 square kilometers for the Chinese empress tree to a staggering 476,298 square kilometers for Cynanchum acutum, a clonal perennial vine known as stranglewort.

The spatial patterns that emerged are striking. Three main invasion hotspots dominate the maps: the Caspian northern lowlands and Alborz foothills in the north, the Persian Gulf and Gulf of Oman coasts in the south, and the Zagros foothills in the southwest. In contrast, the central plateau and the hyperarid basins of Dasht-e Kavir and Dasht-e Lut are unsuitable for most species. Each functional group tells its own story. Salt- and drought-tolerant species such as mesquite, with 279,376 square kilometers of high suitability, dominate the southern coasts and southwestern lowlands. Tree invaders like tree-of-heaven concentrate in the northern provinces along the Caspian coast. Aquatic species, including water lettuce, water hyacinth, and water fern, cluster around the Caspian littoral wetlands, while succulents such as Indian fig opuntia and the halophytic shoreline purslane form nearly continuous bands along the southern shoreline.

When the individual risk maps were overlaid to produce a cumulative picture, the Caspian coastal plain emerged as the country’s most urgent priority, with overlapping extreme- and high-impact zones supporting tree-of-heaven, western ragweed, stranglewort, fiddleneck, and the aquatic invaders. The southern and southeastern regions, including Sistan and Baluchestan and the Zagros foothills, ranked as the second most critical area, driven by mesquite, North Indian rosewood, cogon grass, and valley redstem. Under mid-century climate projections, the number of species contributing to rising risk in the northern coastal regions climbs from four to six under the intermediate RCP 4.5 scenario to seven to nine under the high-emission RCP 8.5 scenario, while the central deserts show declining contributions.

Perhaps the most surprising finding is what did not happen. Rather than a uniform upward shift toward higher impact categories, climate change projections revealed a pronounced bidirectional redistribution among risk classes. Some species expand dramatically in low-risk areas while contracting in moderate and very high zones, as seen with cogon grass, which gained roughly 370,000 square kilometers in low-risk areas while losing ground in higher categories. Stranglewort showed the most dramatic turnover of all, expanding by about 640,000 square kilometers in moderate-risk areas while losing 425,000 square kilometers from very high zones. Aquatic species each gained nearly 300,000 square kilometers in low-impact areas under RCP 8.5 while losing up to 360,000 square kilometers from moderate and high classes. Extreme-risk zones, meanwhile, remained geographically restricted and did not expand proportionally under either scenario, indicating that climate change reshuffles invasion risk as much as it amplifies it.

The researchers are candid about the framework’s limitations. EICAT assessments carry epistemic uncertainty from incomplete ecological knowledge and linguistic uncertainty from standardized terminology, challenges amplified in regions where local impact evidence is scarce. Most of the EICAT scores in this study were drawn from global datasets, meaning local impacts may differ from the global maximum categories used. Species distribution models introduce their own uncertainties, including biased occurrence data, uneven sampling, and pseudo-absence selection. By combining the two tools, RIRC necessarily accumulates uncertainty from both sources. The authors also defended their choice of equal weighting for impact and suitability, arguing that any alternative weighting scheme would embed subjective judgments about which component matters more, and that situations of high impact with low suitability, or the reverse, ultimately require normative decisions about acceptable levels of risk.

Even with these caveats, the practical implications are clear. The findings point to region-specific strategies built around early detection and rapid response, with monitoring prioritized for hydrological corridors in the north, plantations and ports in the south, and sentinel plots at high-elevation boundaries where species like spotted spurge and valley redstem are beginning to climb. At the national level, the framework supports predictive pathway management, climate-informed reprioritization of hotspots, and strategically deployed detection teams. RIRC is presented as an evolving tool rather than a fixed model: as EICAT assessments become more locally informed and distribution models improve with expanded sampling and higher-resolution environmental data, its reliability will grow. Future refinements could integrate land-use dynamics, soil traits, hydrological variables, and anthropogenic disturbances. For now, the framework offers something invasion biology has long needed: a transparent, scalable way to translate global impact assessments into locally actionable maps, before the invaders arrive.

Subject of Research: A predictive framework combining EICAT impact scores with species distribution models to classify regional invasion risks of alien plants in Iran under current and future climates

Article Title: Regional impact risk classification (RIRC): A predictive framework for plant invasion impacts

Article References: Oveisi, M., Hazrati, Z., Sun, Y., Hejazi, M., Bacher, S., & Mueller Schaerer, H. (2026). Regional impact risk classification (RIRC): A predictive framework for plant invasion impacts. NeoBiota, 107, 243-268. https://doi.org/10.3897/neobiota.107.182605

Image Credits: AI Generated

DOI: 10.3897/neobiota.107.182605

Keywords: invasive species, plant invasion, risk classification, EICAT, species distribution models, climate change, Iran, biodiversity, invasion hotspots, NeoBiota, ecological impact, conservation planning

The dataset, described in the Biodiversity Data Journal, covers surveys conducted from March 2007 to June 2023 and documents six mammal taxa legally designated as Endangered Wildlife under Korea’s Wildlife Protection and Management Act: the Eurasian otter (Lutra lutra), the leopard cat (Prionailurus bengalensis), the Siberian flying squirrel subspecies Pteromys volans aluco, the yellow-throated marten (Martes flavigula), the least weasel (Mustela nivalis) and Ognev’s long-eared bat (Plecotus ognevi). Each record represents a detection of one of these animals tied to a specific development project, a specific project phase and a specific survey event, drawn from 73 development projects scattered across the country.

What makes the work technically interesting is not just the volume of records but the provenance. Environmental impact assessment is a decision-support process required before major development proceeds, and its ecological surveys generate species occurrence information that feeds into impact predictions and mitigation planning. Post-EIA monitoring then extends that record through the construction and operation phases of a project, producing repeated surveys over years. Together, these documents collectively contain an enormous volume of biodiversity data — but in Korea, as in many countries, that information has historically been dispersed across individual reports, locked in inconsistent formats and unavailable as structured data that researchers can actually reuse.

To unlock it, the researchers retrieved publicly available EIA and post-EIA monitoring reports through Korea’s national Environmental Impact Assessment Support System. They focused on projects whose EIA consultation process was completed between 2010 and 2019 and whose post-EIA monitoring had been completed by 2024, then filtered for projects in which endangered mammals had actually been documented. Road-construction projects were excluded because records from those projects had already been compiled and published in a previous study. Because assessment-phase surveys can precede the formal completion of the consultation process, some records carry survey dates earlier than 2010, stretching the temporal coverage back to 2007.

The extraction process was painstaking. For each report, the team pulled out species names, survey dates, detection methods and occurrence locations. Crucially, the locations of animal detections were often shown only as points on maps within the reports, so the researchers manually digitised those mapped locations using Google Earth Pro. The resulting coordinates therefore reflect the scale and detail of the source maps rather than original field GPS readings — a limitation the team handled explicitly. To protect sensitive location information for legally protected species, a standard conservation safeguard against poaching and disturbance, the published coordinates were generalised to the centre of the corresponding one-kilometre grid cell, and every record was assigned a coordinate uncertainty of 800 metres to conservatively account for displacement within the cell and positional error from digitisation. The original, non-generalised coordinates are not included in the public dataset.

Detection methods varied widely, and the dataset preserves that diversity. Records are based on direct observations, indirect evidence such as tracks, scat, feeding signs and habitat evidence, camera-trap records and other survey methods reported in the original documents. Each record is classified in the Darwin Core Archive as either a HumanObservation or a MachineObservation depending on how the animal was detected. Records from different projects, project phases or survey events were retained as separate occurrences even when the same species was recorded at the same spot, preserving the temporal depth that makes post-EIA monitoring data so valuable for studying population trends.

Quality control was a multi-stage affair. Species names and taxonomic information were reviewed and standardised against the National Species List of Korea, maintained by the National Institute of Biological Resources. Geographic coordinates were cross-checked against the occurrence locations shown on the source maps, and spatial inconsistencies and data-entry errors were identified and corrected. Duplicate records — defined as repeated records of the same taxon at the same location within the same project, phase and survey event — were consolidated into single occurrences. Survey dates were standardised to year or year-month according to the temporal resolution available in the source reports, and the entire dataset was structured using Darwin Core terms, the community standard for biodiversity occurrence data, before publication through GBIF under a Creative Commons public domain waiver.

The geographic sweep is broad. The records span latitudes from roughly 34.6 to 38.1 degrees north and longitudes from about 126.4 to 129.4 degrees east, covering diverse regions and habitat types across the Republic of Korea — forests, wetlands, rivers, agricultural areas, coastal zones and urban environments. That breadth matters because the six species in the dataset are exactly the kind of animals that development projects threaten most. Mammals with large home ranges and high mobility are particularly vulnerable to habitat loss, fragmentation and impeded movement caused by infrastructure, so spatial occurrence information for these species is directly useful for assessing development impacts and evaluating whether mitigation measures actually work.

The researchers are careful about what the dataset is not. It should not be interpreted as the result of a systematic or comprehensive survey of endangered mammal distributions in Korea; its spatial and temporal coverage reflects only the locations and survey periods of the development projects included in the source reports, and no records from other published sources or external biodiversity databases were added. Every record is marked as a presence, and additional fields record each species’ legal conservation class in Korea, either Class I or Class II Endangered Wildlife. That honesty about scope is part of a broader push in the field: international guidance, including work by the GBIF Secretariat and the International Association for Impact Assessment, recommends that biodiversity records generated through environmental assessment be standardised and made openly accessible, and recent studies in other countries have begun mobilising such records in the same way.

The significance goes beyond Korea’s borders. Environmental impact assessments are legally mandated in most developed economies, which means that decades of ecological survey effort — funded by developers, conducted by consultants, reviewed by regulators — are sitting in document repositories worldwide, invisible to the scientists who could use them. Converting that paper trail into FAIR data, findable, accessible, interoperable and reusable, turns a compliance burden into a research asset. For Korea’s endangered mammals, the 5,000 records now openly available represent sixteen years of detection history across 73 projects, offering conservation planners and researchers a baseline for distribution studies that would otherwise have cost millions of won in new fieldwork. The dataset is available through GBIF as a Darwin Core Archive, and the underlying data paper documents every step of the compilation, from map digitisation to coordinate generalisation, so that others can replicate the approach in their own countries.

Subject of Research: Compilation of endangered mammal occurrence records from environmental impact assessment and post-EIA monitoring reports in the Republic of Korea

Article Title: Endangered mammal occurrence records from Environmental Impact Assessment and post-EIA monitoring in the Republic of Korea

Article References: Kim, J., Kim, H.-R., Chang, M., & Kim, N. (2026). Endangered mammal occurrence records from Environmental Impact Assessment and post-EIA monitoring in the Republic of Korea. Biodiversity Data Journal, 14, Article e212750. https://doi.org/10.3897/bdj.14.e212750

Image Credits: AI Generated

DOI: 10.3897/BDJ.14.e212750

Keywords: environmental impact assessment, biodiversity data, endangered mammals, GBIF, Darwin Core, South Korea, wildlife monitoring, species occurrence records, conservation planning, National Institute of Ecology, post-EIA monitoring, open data

The research team, working within the Central Visayas region of the Philippines, faced a familiar dilemma in tropical conservation: the species most in need of protection are often the ones about which least is known. Many of the threatened endemics in Rajah Sikatuna have sparse, scattered occurrence records, making it difficult to decide which forest patches deserve the strictest protection. To overcome this, the researchers turned to Species Distribution Modelling, a technique that links known locations of a species with environmental variables to predict where suitable habitat may exist across the wider landscape. Using the MaxEnt algorithm, they modelled habitat suitability for thirteen threatened species spanning birds, mammals, amphibians, reptiles and trees.

The environmental backbone of the models came from two complementary sources. Nineteen bioclimatic variables, describing temperature and precipitation patterns, were interpolated from observations provided by the Philippine Atmospheric, Geophysical and Atmospheric Services Administration using inverse distance weighting, a deterministic method chosen deliberately because only three local sampling points were available, too few to fit the semi-variogram models required by geostatistical approaches such as kriging. Terrain variables, including elevation, slope, aspect and the Topographic Ruggedness Index, were derived from a 12.5-metre resolution digital elevation model supplied by the ALOS satellite mission, fine enough to capture the rugged karst topography that shapes microclimates and species movements across the reserve.

Species occurrence records were compiled from the Global Biodiversity Information Facility, supplemented by local biodiversity monitoring reports from the Department of Environment and Natural Resources and field surveys conducted by the Central Visayas Agriculture, Aquatic and Natural Resources Research and Development Consortium. After cleaning for duplicate entries and erroneous coordinates, and applying spatial thinning to reduce sampling bias, the final dataset contained just eighty-three spatially independent records across all thirteen species. To guard against overfitting with such limited data, the researchers reduced an initial set of twenty-three environmental predictors to five statistically independent ones, all with variance inflation factors below four, well under the conservative threshold of five: Annual Mean Temperature, Annual Precipitation, Elevation, Slope and Terrain Ruggedness Index.

The resulting models performed remarkably well for a data-poor tropical landscape. The mean Area Under the Curve, a standard measure of a model’s ability to discriminate between suitable and unsuitable habitat, reached 0.905, with individual species scores ranging from 0.860 for the Mindanao cuckooshrike Edolisoma mindanense to 0.950 for the critically endangered Mindanao bleeding-heart Gallicolumba crinigera. The mean True Skill Statistic, which measures accuracy independently of how common a species is, stood at 0.717, comfortably above the 0.60 benchmark generally regarded as indicating substantial model performance. Validation used five-fold cross-validation for species with fewer than ten records and a leave-one-out jackknife procedure for the most data-limited taxa, ensuring that performance estimates remained reliable across varying sample sizes.

One of the study’s most technically careful decisions concerned how continuous probability surfaces were converted into binary presence-absence maps. The researchers initially considered the minimum training presence threshold, which minimises omission errors, but a sensitivity analysis revealed that this approach inflated predicted presence by up to seventy-four percent in certain taxa, effectively encoding spatial noise into the conservation maps. They instead adopted the maximum sum of sensitivity and specificity threshold, widely recognised as one of the most robust options for conservation planning because it optimises the balance between failing to detect suitable habitat and wrongly classifying unsuitable areas as suitable.

The maps that emerged revealed a landscape of stark ecological contrasts. Some species, including the Mindanao bleeding-heart, the dipterocarp tree Dipterocarpus grandiflorus and the hardwood Shorea guiso, showed highly restricted and fragmented suitability patterns, concentrated in the western and northern sectors where intact forest interiors persist. Others, such as the sailfin lizard Hydrosaurus pustulatus, the long-tailed macaque Macaca fascicularis and the winchell’s kingfisher Todiramphus winchelli, occupied narrow localised hotspots vulnerable to site-specific disturbance. In contrast, the amphibian Platymantis guentheri, the cuckooshrike and the tree Voacanga megacarpa displayed broader, more continuous suitability gradients, suggesting greater ecological plasticity and resilience to moderate landscape-level change.

The environmental drivers behind these patterns were equally species-specific. Annual Mean Temperature contributed 29.2 percent and Elevation 28.2 percent across the ensemble, but individual models told sharper stories: temperature dominated the model for the Mindanao bleeding-heart at 62.3 percent, while Annual Precipitation drove sixty percent of the model for the karst-dwelling frog Platymantis guentheri, reflecting the physiological dependence of amphibians on moisture regimes. Elevation was the primary predictor for dipterocarp trees, contributing 54.9 percent for Pitta steerii and 45.2 percent for Dipterocarpus grandiflorus, while terrain ruggedness and slope shaped the distributions of the macaque and the bent-winged bat Miniopterus schreibersii respectively. The researchers are careful to note that these climatic drivers, interpolated from only three sampling points, should be read as proxy gradients reflecting broad landscape transitions rather than precise mechanistic limits on where species can live.

Aggregating the individual binary maps produced a species richness map that identified the northern and north-western sections of the reserve, with richness classes of ten to twelve species, as the primary biodiversity hotspots, corresponding to areas of relatively intact forest cover and topographic complexity. The southern and eastern portions, classified in the lowest richness classes of zero to three species, likely reflect degraded or fragmented habitats offering fewer niches and fewer resources. The frequency distribution of richness across the landscape was multimodal, with a primary peak at three species per pixel and a secondary peak at five, confirming that the landscape is dominated by moderate biodiversity while high-richness areas remain spatially rare and restricted.

The final and arguably most consequential step was deciding where to draw the line between protected core zones and the wider landscape. The researchers tested five statistical thresholds against the richness distribution, measuring how many known species occurrences each captured. The most conservative options, isolating only the top five percent of the landscape, captured just twenty-nine percent of occurrences, while the ninetieth and ninety-fifth percentiles pinpointed rare hotspots but excluded moderately rich areas that may function as corridors or buffers. The winning candidate was the inflection point of the Cumulative Richness Frequency Distribution, at approximately six species per cell, which captured ninety percent of all species occurrences while maintaining a spatially inclusive footprint. Because Rajah Sikatuna already holds formal protected status, the authors argue that thresholds must balance ecological significance with management feasibility, and the inflection-point approach achieves exactly that, protecting both primary biodiversity nodes and the functional buffer habitats required for long-term species persistence. The team frames its outputs not as immutable boundaries for legal zoning but as spatial hypotheses and decision-support tools, intended to prioritise future ground-truthing, guide denser sensor deployment and steer restoration efforts toward the degraded southern slopes where recovery could reconnect fragmented habitat islands.

Subject of Research: Species distribution modelling and conservation prioritisation of threatened species in a Philippine protected landscape

Article Title: Identification of potential suitable habitats and high conservation value areas for threatened species in Rajah Sikatuna Protected Landscape, Philippines

Article References: Aureo, W., Tiburan Jr., C., Calzeta, E., Lit Jr., I., Yan, Y., Taylor, D. M., & Ancog, R. (2026). Identification of potential suitable habitats and high conservation value areas for threatened species in Rajah Sikatuna Protected Landscape, Philippines. One Ecosystem, 11, Article e184595. https://doi.org/10.3897/oneeco.11.e184595

Image Credits: AI Generated

DOI: 10.3897/oneeco.11.e184595

Keywords: species distribution modelling, MaxEnt, Rajah Sikatuna Protected Landscape, Philippines, threatened species, High Conservation Value Areas, species richness, karst forest, habitat suitability, conservation planning, biodiversity hotspot, endemic species