Predictive Species Distribution Models Research Articles

Applying landscape metrics to species distribution model predictions to characterize internal range structure and associated changes.

Distributional shifts in species ranges provide critical evidence of ecological responses to climate change. Assessments of climate-driven changes typically focus on broad-scale range shifts (e.g. poleward or upward), with ecological consequences at regional and local scales commonly overlooked. While these changes are informative for species presenting continuous geographic ranges, many species have discontinuous distributions-both natural (e.g. mountain or coastal species) or human-induced (e.g. species inhabiting fragmented landscapes)-where within-range changes can be significant. Here, we use an ecosystem engineer species (Sabellaria alveolata) with a naturally fragmented distribution as a case study to assess climate-driven changes in within-range occupancy across its entire global distribution. To this end, we applied landscape ecology metrics to outputs from species distribution modelling (SDM) in a novel unified framework. SDM predicted a 27.5% overall increase in the area of potentially suitable habitat under RCP 4.5 by 2050, which taken in isolation would have led to the classification of the species as a climate change winner. SDM further revealed that the latitudinal range is predicted to shrink because of decreased habitat suitability in the equatorward part of the range, not compensated by a poleward expansion. The use of landscape ecology metrics provided additional insights by identifying regions that are predicted to become increasingly fragmented in the future, potentially increasing extirpation risk by jeopardising metapopulation dynamics. This increased range fragmentation could have dramatic consequences for ecosystem structure and functioning. Importantly, the proposed framework-which brings together SDM and landscape metrics-can be widely used to study currently overlooked climate-driven changes in species internal range structure, without requiring detailed empirical knowledge of the modelled species. This approach represents an important advancement beyond predictive envelope approaches and could reveal itself as paramount for managers whose spatial scale of action usually ranges from local to regional.

Beyond topo-climatic predictors: Does habitats distribution and remote sensing information improve predictions of species distribution models?

Species Distribution Models (SDM) represent a powerful tool to predict species’ habitat suitability on a landscape and fill the gap between truncated observation data and all possible locations. SDMs have been widely used in theoretical studies of species niches as well as in conservation applications. Here, we evaluated the impacts of predictors’ type on models’ performances and spatial predictions using 72 plant species belonging to six ecological groups at a regional scale in the area of Geneva (Switzerland). Twelve models were created using various combinations of high-resolution (25 m) explanatory variables including topography, pedology, climate, habitats and remote sensing data. Models integrating a combination of habitats and topopedo-climatic predictors had significantly higher performances, while remote sensing predictors showed low performances. Our results suggest that the number and the level of details of habitat predictors (broad or very precise) do not fundamentally affect prediction maps. However, selecting too few, overly simplified or exceedingly complex habitat predictors tend to lower models’ performances. The use of eight habitat categories complemented with eight topopedo-climatic predictors produced models with the highest performances. Ecological groups of species responded differently to models and while alpine and ruderal species have greater average performances due to a high affinity with topopedo-climatic predictors, wetlands’ species were less performant on average. These results underline the necessity of developing or having access to habitats distribution data especially in a conservation context. • Models with biotic and topo-climatic predictors show the highest performances. • Habitat predictors characteristics have little impact on prediction maps. • When used raw, NDVI and NDWI do not replace biotic information. • Ecological groups of species responded differently to models. • The creation of LULC maps ensure better conservation practices.

Modeling the geographical distributions of Chordodes formosanus and its mantis hosts in Taiwan, with considerations for their niche overlaps.

Species distribution models (SDMs) have conventionally been used for evaluating the distribution of individual species, but they can also be used, through comparing different SDMs, to evaluate the geographic similarity between taxa. In this study, we used a parasite and host system to infer the geographic overlaps between species with tight biological interaction, for example, parasites and their obligate host. Specifically, we used the horsehair worm Chordodes formosanus and its three mantis hosts to study the extent of niche overlap. We retrieved presence points for the host species and the parasite, and then we built SDMs with MaxEnt implemented in ENMeval using selected bioclim variables (based on variance inflation factor values) at 30s scale. The models showed that the hosts and parasite do not occur in the high elevation areas in Taiwan, which is expected based on their biology. Interestingly, the predicted parasite distribution included areas without collection records, implying local extinction or sampling bias. We subsequently evaluated niche overlap between hosts and the parasite according to five similarity indices (Schoener's D, I statistic, relative rank, Pearson correlation coefficient, and the rank correlation coefficient rho). Our models showed a high similarity of SDM predictions between hosts and the parasite. There were differences among metrics for which host shared the highest similarity with the parasite, but the majority of the results indicated that the Japanese boxing mantis had the highest niche similarity with the horsehair worm. The choice of the niche overlap metric to use can uncover information on the parasite's ecology, which can be important for endangered species. SDMs are reliable tools for host and parasite conservation management and could help improve our understanding of parasite biology and ecology.

How far will they go? Considerations on the potential expansion of the Ring-necked Parakeet (<em>Psittacula krameri</em>) and Monk Parakeet (<em>Myiopsitta monachus</em>) in Veneto region (Italy) with MaxEnt distribution models

Predictive models of species distribution, including several parakeets, may be very useful for understanding the actual and potential distribution of alien species. The Ring-necked Parakeet and the Monk Parakeet are two alien species found in the Veneto region (north-east Italy) that are well-suited for conducting potential distribution studies to identify suitable areas of occurrence. In this paper, I use the MAXENT algorithm to predict the potential distribution maps of these species in the Veneto region, which finds the probability distribution of maximum entropy that is constrained by the ecological parameters considered. The authors of the Veneto photo-sound atlas kindly provided presence data for the two species. The potential distribution models for the two parakeets performed very well overall (AUC = 0.966 for Ring-necked Parakeet and AUC = 0.987 for Monk Parakeet) and indicated that high suitability areas correspond primarily to areas near known occurrences of the species. Overall, the study found that the presence of the species in provinces other than those with large populations is highly unlikely. The distribution maps produced can aid in the selection of monitoring areas for these two alien species’ populations and potential expansion.

Impact of climate change on potential distribution of Quercus suber in the conditions of North Africa

Climate change, which is expected to continue in the future, is increasingly becoming a major concern affecting many components of the biodiversity and human society. Understanding its impacts on forest ecosystems is essential for undertaking long-term management and conservation strategies. This study was focused on modeling the potential distribution of Quercus suber in the Maamora Forest, the world’s largest lowland cork oak forest, under actual and future climate conditions and identifying the environmental factors associated with this distribution. Maximum Entropy approach was used to train a Species Distribution Model and future predictions were based on different greenhouse gas emission scenarios (Representative Concentration Pathway RCPs). The results showed that the trained model was highly reliable and reflected the actual and future distributions of Maamora’s cork oak. It showed that the precipitation of the coldest and wettest quarter and the annual temperature range are the environmental factors that provide the most useful information for Q. suber distribution in the study area. The computed results of cork oak’s habitat suitability showed that predicted suitable areas are site-specific and seem to be highly dependent on climate change. The predicted changes are significant and expected to vary (decline of habitat suitability) in the future under the different emissions pathways. It indicates that climate change may reduce the suitable area for Q. suber under all the climate scenarios and the severity of projected impacts is closely linked to the magnitude of the climate change. The percent variation in habitat suitability indicates negative values for all the scenarios, ranging –23% to –100%. These regressions are projected to be more important under pessimist scenario RCP8.5. Given these results, we recommend including the future climate scenarios in the existing management strategies and highlight the usefulness of the produced predictive suitability maps under actual and future climate for the protection of this sensitive forest and its key species – cork oak, as well as for other forest species.

Transferable species distribution modelling: Comparative performance of Generalised Functional Response models

Predictive species distribution models (SDMs) are becoming increasingly important in ecology, in the light of rapid environmental change. However, the predictions of most current SDMs are specific to the habitat composition of the environments in which they were fitted. This may limit SDM predictive power because species may respond differently to a given habitat depending on the availability of all habitats in their environment, a phenomenon known as a functional response in resource selection. The Generalised Functional Response (GFR) framework captures this dependence by formulating the SDM coefficients as functions of habitat availability. The original GFR implementation used global polynomial functions of habitat availability to describe the functional responses. In this study, we develop several refinements of this approach and compare their predictive performance using two simulated and two real datasets. We first use local radial basis functions (RBF), a more flexible approach than global polynomials, to represent the habitat selection coefficients, and balance bias with precision via regularization to prevent overfitting. Second, we use the RBF-GFR and GFR models in combination with the classification and regression tree CART, which has more flexibility and better predictive powers for non-linear modelling. As further extensions, we use random forests (RFs) and extreme gradient boosting (XGBoost), ensemble approaches that consistently lead to variance reduction in generalization error. We find that the different methods are ranked consistently across the datasets for out-of-data prediction. The traditional stationary approach to SDMs and the GFR model consistently perform at the bottom of the ranking (simple SDMs underfit, and polynomial GFRs overfit the data). The best methods in our list provide non-negligible improvements in predictive performance, in some cases taking the out-of-sample R2 from 0.3 up to 0.7 across datasets. At times of rapid environmental change and spatial non-stationarity ignoring the effects of functional responses on SDMs, results in two different types of prediction bias (under-prediction or mis-positioning of distribution hotspots). However, not all functional response models perform equally well. The more volatile polynomial GFR models can generate biases through over-prediction. Our results indicate that there are consistently robust GFR approaches that achieve impressive gains in transferability across very different datasets.

Bat responses to climate change: a systematic review.

Understanding how species respond to climate change is key to informing vulnerability assessments and designing effective conservation strategies, yet research efforts on wildlife responses to climate change fail to deliver a representative overview due to inherent biases. Bats are a species-rich, globally distributed group of organisms that are thought to be particularly sensitive to the effects of climate change because of their high surface-to-volume ratios and low reproductive rates. We systematically reviewed the literature on bat responses to climate change to provide an overview of the current state of knowledge, identify research gaps and biases and highlight future research needs. We found that studies are geographically biased towards Europe, North America and Australia, and temperate and Mediterranean biomes, thus missing a substantial proportion of bat diversity and thermal responses. Less than half of the published studies provide concrete evidence for bat responses to climate change. For over a third of studied bat species, response evidence is only based on predictive species distribution models. Consequently, the most frequently reported responses involve range shifts (57% of species) and changes in patterns of species diversity (26%). Bats showed a variety of responses, including both positive (e.g. range expansion and population increase) and negative responses (range contraction and population decrease), although responses to extreme events were always negative or neutral. Spatial responses varied in their outcome and across families, with almost all taxonomic groups featuring both range expansions and contractions, while demographic responses were strongly biased towards negative outcomes, particularly among Pteropodidae and Molossidae. The commonly used correlative modelling approaches can be applied to many species, but do not provide mechanistic insight into behavioural, physiological, phenological or genetic responses. There was a paucity of experimental studies (26%), and only a small proportion of the 396 bat species covered in the examined studies were studied using long-term and/or experimental approaches (11%), even though they are more informative about the effects of climate change. We emphasise the need for more empirical studies to unravel the multifaceted nature of bats' responses to climate change and the need for standardised study designs that will enable synthesis and meta-analysis of the literature. Finally, we stress the importance of overcoming geographic and taxonomic disparities through strengthening research capacity in the Global South to provide a more comprehensive view of terrestrial biodiversity responses to climate change.

Biotic predictors improve species distribution models for invasive plants in Western U.S. Forests at high but not low spatial resolutions

Invasions by non-native plants threaten forest health and sustainability. The ability to predict areas at greatest risk to invasion is essential for informing both monitoring and management of invasive species. Species distribution models (SDMs) are often used to identify environmental correlates of a species’ occurrence and predict geographic areas that may be suitable for its presence and are commonly constructed using solely abiotic predictors. However, mounting evidence implies that not including biotic predictors in SDMs may yield less accurate models at some resolutions typical of landscape-scale models, although this possibility has rarely been evaluated in invasive plants. In this study, we determined whether including descriptors of the biotic environment improved the accuracy of SDMs built at five decreasing spatial resolutions for infestations of five common invasive plants in forests of California, Oregon, and Washington, USA and described environmental correlates of each species’ presence.Predictors of occurrence often echoed those identified in previous studies of the focal species, indicating that our models accurately identified important environmental drivers of occurrence. Including biotic predictors in the SDMs consistently improved model accuracy only at the highest resolution we examined, which may be due to the spatial scale at which biotic interactions primarily act to constrain species’ distributions, the particular biotic predictors we used in our models, or correlations between attributes of the abiotic and biotic environment. This finding suggests that, while the practice of building SDMs using abiotic predictors alone may generally yield models whose accuracy does not differ substantially from those that also include biotic predictors, the effects of biotic interactions on the distribution of invasive plants in forests may be detectable at larger scales than previously thought.

Effects of input data sources on species distribution model predictions across species with different distributional ranges

AbstractAimA major source of uncertainty in the application of species distribution models (SDMs) is related to input data quality. Citizen‐collected species occurrence data are often used for fitting SDMs when data from standardized and expert‐supported surveys are unavailable. Macroclimate variables are much more commonly used as predictors in SDMs than other sources coming from remote sensing data. Here, we assess the effects of using different data sources (in both response and predictor variables) on SDM performance across a wide range of species with contrasting distributional ranges.LocationIberian Peninsula.TaxonBirds.MethodsA SDM ensemble‐forecasting approach was implemented using bird data from two different data sources: the eBird project and Atlases. We fitted SDMs with three predictor types: macroclimate, remotely sensed ecosystem functional attributes (EFAs) and their combination. Species were grouped in four range size classes. We assessed the uncertainty of model predictions by different evaluation metrics. Generalized linear mixed‐effects models tested the effect on model performance of input data source across distributional range sizes while accounting for different accuracy metrics. Pairwise comparisons between range projections were used to assess their spatial similarity.ResultsData source, size class, predictor and accuracy metric showed significant effects on SDM performance. eBird‐based models outperformed those built with Atlas data for less widespread species. Climate predictors yielded models with the best performance, especially when combined with EFAs. However, the predictor contribution was consistent across bird datasets, being mostly driven by the species range.Main ConclusionsOur models demonstrated the usefulness and complementarity of different input data sources when modelling species distribution across different distributional ranges. These findings highlight the need to integrate different data sources to improve the model predictions at regional scale. Our framework also underlines that model uncertainty should be examined more exhaustively at early stages of the modelling process.

Habitats as predictors in species distribution models: Shall we use continuous or binary data?

The representation of a land cover type (i.e. habitat) within an area is often used as an explanatory variable in species distribution models. However, it is possible that a simple binary presence/absence of the suitable habitat might be the most important determinant of the presence/absence of some species and, thus, be a better predictor of species occurrence than the continuous parameter (area). We hypothesize that the binary predictor is more suitable for relatively rare habitats (e.g. wetlands) while for common habitats (e.g. forests) the amount of the focal habitat is a better predictor. We used the Third Atlas of Breeding Birds in the Czech Republic as the source of species distribution data and CORINE Land Cover inventory as the source of the landcover information. To test our hypothesis, we fitted generalized linear models of 32 water and 32 forest bird species. Our results show that for water bird species, models using binary predictors (presence/absence of the habitat) performed better than models with continuous predictors (i.e. the amount of the habitat); for forest species, however, we observed the opposite. Thus, future studies using habitats as predictors of species occurrences should consider the prevalence of the habitat in the landscape, and the biological role of the habitat type in the particular species' life history. In addition, performing a preliminary comparison of the performance of the binary and continuous versions of habitat predictors (e.g. using information criteria) prior to modelling, during variable selection, can be beneficial. These are simple steps that will improve explanatory and predictive performance of models of species distributions in biogeography, community ecology, macroecology and ecological conservation.

Leaf morphological traits as adaptations to multiple climate gradients.

Leaf morphological traits vary systematically along climatic gradients. However, recent studies in plant functional ecology have mainly analysed quantitative traits, while numerical models of species distributions and vegetation function have focused on traits associated with resource acquisition; both ignore the wider functional significance of leaf morphology.A dataset comprising 22 leaf morphological traits for 662 woody species from 92 sites, representing all biomes present in China, was subjected to multivariate analysis in order to identify leading dimensions of trait covariation (correspondence analysis), quantify climatic and phylogenetic contributions (canonical correspondence analysis with variation partitioning) and characterise co‐occurring trait syndromes (k‐means clustering) and their climatic preferences.Three axes accounted for >20% of trait variation in both evergreen and deciduous species. Moisture index, precipitation seasonality and growing‐season temperature explained 8%–10% of trait variation; family 15%–32%. Microphyll or larger, mid‐ to dark green leaves with drip tips in wetter climates contrasted with nanophyll or smaller glaucous leaves without drip tips in drier climates. Thick, entire leaves in less seasonal climates contrasted with thin, marginal dissected, aromatic and involute/revolute leaves in more seasonal climates. Thick, involute, hairy leaves in colder climates contrasted with thin leaves with marked surface structures (surface patterning) in warmer climates. Distinctive trait clusters were linked to the driest and most seasonal climates, for example the clustering of picophyll, fleshy and succulent leaves in the driest climates and leptophyll, linear, dissected, revolute or involute and aromatic leaves in regions with highly seasonal rainfall. Several trait clusters co‐occurred in wetter climates, including clusters characterised by microphyll, moderately thick, patent and entire leaves or notophyll, waxy, dark green leaves. Synthesis. The plastic response of size, shape, colour and other leaf morphological traits to climate is muted, thus their apparent shift along climate gradients reflects plant adaptations to environment at a community level as determined by species replacement. Information on leaf morphological traits, widely available in floras, could be used to strengthen predictive models of species distribution and vegetation function.

Species distribution models as a tool for early detection of the invasive Raphidiopsis raciborskii in European lakes

In freshwater habitats, invasive species and the increase of cyanobacterial blooms have been identified as a major cause of biodiversity loss. The invasive cyanobacteria Raphidiopsis raciborskii a toxin-producing and bloom-forming species affecting local biodiversity and ecosystem services is currently expanding its range across Europe. We used species distribution models (SDMs) and regional bioclimatic environmental variables, such as temperature and precipitation, to identify suitable areas for the colonization and survival of R. raciborskii, with special focus on the geographic extent of potential habitats in Northern Europe. SDMs predictions uncovered areas of high occurrence probability of R. raciborskii in locations where it has not been recorded yet, e.g. some areas in Central and Northern Europe. In the southeastern part of Sweden, areas of suitable climate for R. raciborskii corresponded with lakes of high concentrations of total phosphorus, increasing the risk of the species to thrive. To our knowledge, this is the first attempt to predict areas at high risk of R. raciborskii colonization in Europe. The results from this study suggest several areas across Europe that would need monitoring programs to determine if the species is present or not, to be able to prevent its potential colonization and population growth. Regarding an undesirable microorganism like R. raciborskii, authorities may need to start information campaigns to avoid or minimize the spread.

The future distribution of wetland birds breeding in Europe validated against observed changes in distribution

Wetland bird species have been declining in population size worldwide as climate warming and land-use change affect their suitable habitats. We used species distribution models (SDMs) to predict changes in range dynamics for 64 non-passerine wetland birds breeding in Europe, including range size, position of centroid, and margins. We fitted the SDMs with data collected for the first European Breeding Bird Atlas and climate and land-use data to predict distributional changes over a century (the 1970s–2070s). The predicted annual changes were then compared to observed annual changes in range size and range centroid over a time period of 30 years using data from the second European Breeding Bird Atlas. Our models successfully predicted ca. 75% of the 64 bird species to contract their breeding range in the future, while the remaining species (mostly southerly breeding species) were predicted to expand their breeding ranges northward. The northern margins of southerly species and southern margins of northerly species, both, predicted to shift northward. Predicted changes in range size and shifts in range centroids were broadly positively associated with the observed changes, although some species deviated markedly from the predictions. The predicted average shift in core distributions was ca. 5 km yr−1 towards the north (5% northeast, 45% north, and 40% northwest), compared to a slower observed average shift of ca. 3.9 km yr−1. Predicted changes in range centroids were generally larger than observed changes, which suggests that bird distribution changes may lag behind environmental changes leading to ‘climate debt’. We suggest that predictions of SDMs should be viewed as qualitative rather than quantitative outcomes, indicating that care should be taken concerning single species. Still, our results highlight the urgent need for management actions such as wetland creation and restoration to improve wetland birds’ resilience to the expected environmental changes in the future.

Black Stork Back: Species distribution model predictions of potential habitats for Black Stork Ciconia nigra in Sweden

Increased understanding of the need to save endangered and locally extinct species has led to restoration or preservation of populations through reintroductions. Reintroduction of a species is worthwhile if the prerequisites for existence at the historical location have improved. Thus, background information about the habitat requirements of a target species is important for introduction programmes to be successful. The Black Stork Ciconia nigra was lost as a breeding species in Sweden during the 20th century, but recent observations and reports of potential breeding indicate that habitat conditions for Black Stork in Sweden may have improved. In this study, we used species characteristics and references to identify habitats in Sweden suitable for potential reintroduction of Black Stork. We identified several suitable areas in the former distribution range of this species in southern Sweden. Seven Swedish counties contained more than 18 % suitable habitat within their total area, with highest proportions in Jönköping County (25.8 %), Blekinge County (23.9 %), Västra Götaland County (22.1 %) and Kronoberg County (20.7 %). We suggest these areas to be made the primary targets for Black Stork reintroduction in Sweden.

Assessing the uncertainty arising from standard land‐cover mapping procedures when modelling species distributions

AbstractAimHabitat loss and degradation is one of the main threats to biodiversity worldwide. Ecological modelling community has been increasingly incorporating habitat changes in the species distribution models (SDMs). However, the effect that the uncertainty arising from the standard procedures of land‐use/land‐cover (LULC) mapping can have on SDMs has been overlooked. Remote sensing offers a great source of information to assess the habitat changes at different spatiotemporal scales. However, despite the great variety of satellite imagery preprocessing and supervised classification procedures currently available, their effects on SDMs remain largely unknown.LocationBaixa Limia–Serra do Xurés Natural Park, NW Spain.MethodsWe developed SDMs for 27 bird species with 4 modelling algorithms using 12 sets of the LULC predictors derived from the combination of 3 image preprocessing levels and 4 supervised classification algorithms.ResultsThe preprocessing level, classification method and modelling algorithm had a significant impact on the SDM performance and on the spatial and temporal predictions, with the considerable variation between species. Even if the effects of satellite image processing and classification procedures on the performance and transferability of SDMs were lower than that of the modelling algorithms themselves, the spatial and temporal predictions of SDMs were overall strongly affected by these LULC mapping procedures, sometimes overriding the effect of modelling algorithms.Main conclusionsWe argue that the uncertainty arising from the standard LULC mapping procedures when predicting species distribution dynamics should receive the same level of attention as the variability due to the different types of climate simulations and models often used in the ecological research. We also highlight the need to report accuracy of the satellite‐based LULC maps for each LULC class since some of these classes can be of key importance for certain types of the ecological applications.

Microclimate‐based species distribution models in complex forested terrain indicate widespread cryptic refugia under climate change

AbstractAimSpecies’ climatic niches may be poorly predicted by regional climate estimates used in species distribution models (SDMs) due to microclimatic buffering of local conditions. Here, we compare SDMs generated using a locally validated below‐canopy microclimate model to those based on interpolated weather station data at two spatial scales to determine the effects of scale and topography on potential future below‐canopy warming and species distributions.LocationGreat Smoky Mountains National Park (2,090 km2; NC, TN, USA).Time period1970–2006, late‐century warming.Major taxa studiedVascular plant species of Southern Appalachian forests.MethodsWe compared the fit and spatio‐temporal predictions of SDMs generated using occurrence records of 154 plant species and three climate models: macroclimate (1 km, WorldClim), downscaled climate (based on a 30‐m digital elevation model), and fine‐scale microclimate (30 m) from a below‐canopy sensor network.ResultsWe found that, although SDM fit was similar across models, microclimate‐derived SDMs predicted substantially greater species persistence with 4°C of regional warming, with a difference of 50% of the species pool in some areas. Microclimate models predicted that warming trends will be buffered in high‐elevation and near‐stream habitats. Microclimate SDMs predicted higher stability of mid‐elevation species, particularly in thermally buffered areas near streams, and critically, less change in species composition at high elevation. In contrast, predictions of macroclimate and downscaled climate models were similar despite improved resolution.Main conclusionsOur results demonstrate that careful selection of climate drivers, including local near‐ground validation rather than downscaling solely with elevation, is critical for projecting distributions. They also suggest that some species at risk from climate change might persist, even with 4°C of macroclimate warming, in cryptic refugia buffered by microclimate, pointing to the roles of forest cover and topography in explaining slower‐than‐expected changes in understorey communities. However, certain species, such as those currently occurring on low‐elevation ridges that are sensitive to atmospheric changes, may be at more risk than macroclimate or downscaled climate SDMs suggest.

Phylogenetic analysis of combined mitochondrial genome and 32 nuclear genes provides key insights into molecular systematics and historical biogeography of Asian warty newts of the genus Paramesotriton (Caudata: Salamandridae).

The Paramesotriton Chang, 1935 genus of Asian warty newts is the second most diverse genus in the family Salamandridae, currently containing 14 recognized species from northern Vietnam to southwest-central and southern China. Although species of this genus have been included in previous phylogenetic studies, the origin and interspecific relationships of the genus are still not fully resolved, especially at key nodes in the phylogeny. In this study, we sequenced mitochondrial genomes and 32 nuclear genes from 27 samples belonging to 14 species to reconstruct the interspecific phylogenetic relationships within Paramesotriton and explore its historical biogeography in southern China. Both Bayesian inference and maximum-likelihood analyses highly supported the monophyly of Paramesotriton and its two recognized species groups (P. caudopunctatus and P. chinensis groups) and further identified five hypothetical phylogenetic cryptic species. Biogeographic analyses indicated that Paramesotriton originated in southwestern China (Yunnan-Guizhou Plateau/South China) during the late Oligocene. The time of origin of Paramesotriton corresponded to the second uplift of the Himalayan/Qinghai-Xizang (Tibetan) Plateau (QTP), rapid lateral extrusion of Indochina, and formation of karst landscapes in southwestern China. Principal component analysis (PCA), independent sample t-tests, and niche differentiation using bioclimatic variables based on locations of occurrence suggested that Paramesotriton habitat conditions in the three current regions (West, South, and East) differ significantly, with different levels of climatic niche differentiation. Species distribution model (SDM) predictions indicated that the most suitable distribution areas for the P. caudopunctatus and P. chinensis species groups are western and southern/eastern areas of southern China. This study increases our knowledge of the taxonomy, biodiversity, origin, and suitable distribution areas of the genus Paramesotriton based on phylogenetic, biogeographic, and species distribution models.

Testing the Influence of Seascape Connectivity on Marine-Based Species Distribution Models

Species distribution models (SDMs) are commonly used in ecology to predict species occurrence probability and how species are geographically distributed. Here, we propose innovative predictive factors to efficiently integrate information on connectivity into SDMs, a key element of population dynamics strongly influencing how species are distributed across seascapes. We also quantify the influence of species-specific connectivity estimates (i.e., larval dispersal vs. adult movement) on the marine-based SDMs outcomes. For illustration, seascape connectivity was modeled for two common, yet contrasting, marine species occurring in southeast Australian waters, the purple sea urchin, Heliocidaris erythrogramma, and the Australasian snapper, Chrysophrys auratus. Our models illustrate how different species-specific larval dispersal and adult movement can be efficiently accommodated. We used network-based centrality metrics to compute patch-level importance values and include these metrics in the group of predictors of correlative SDMs. We employed boosted regression trees (BRT) to fit our models, calculating the predictive performance, comparing spatial predictions and evaluating the relative influence of connectivity-based metrics among other predictors. Network-based metrics provide a flexible tool to quantify seascape connectivity that can be efficiently incorporated into SDMs. Connectivity across larval and adult stages was found to contribute to SDMs predictions and model performance was not negatively influenced from including these connectivity measures. Degree centrality, quantifying incoming and outgoing connections with habitat patches, was the most influential centrality metric. Pairwise interactions between predictors revealed that the species were predominantly found around hubs of connectivity and in warm, high-oxygenated, shallow waters. Additional research is needed to quantify the complex role that habitat network structure and temporal dynamics may have on SDM spatial predictions and explanatory power.

Species distribution modeling that overlooks intraspecific variation is inadequate for proper conservation of marula (Sclerocarya birrea, Anacardiaceae)

Species distribution models (SDMs) are often used to predict the impact of climate change on the future distribution of species. However, the application of species-level SDMs often ignores intraspecific variation and can be problematic. The aim of the study was to incorporate intraspecific information into SDMs and compare predictions of subspecies- and species-level SDMs in a widespread African tree species, Sclerocarya birrea (marula). We used bioclimatic variables and terrain roughness index to calibrate nine algorithms in ‘biomod2’ platform. Ensemble models were built to predict the current and future suitable habitat of marula and separately its three subspecies, birrea , caffra and multifoliolata . Projections were made to 2041 − 2060 and 2081 − 2100 under three general circulation models and two shared socio-economic pathways. Results showed an expansion of the suitable habitat of marula as well as subspp. birrea and caffra in the two future time periods. The suitable habitat of subsp. multifoliolata , in contrast, contracted in future. Our results show that species-level SDMs may fail to detect climate change risks for intraspecific taxa, and consequently leading to under- or overestimation of the impacts of climate change. Incorporation of intraspecific variation into SDMs improves predictions of the impact of climate change and helps to identify appropriate conservation and management options. • SDMs that exclude intraspecific information are inadequate in Sclerocarya birrea . • Impact of climate change on subsp. multifoliolata is masked in a species-level SDM. • Intraspecific-level SDMs show robust performance than species-level SDMs. • Conservation should be based on predictions of intraspecific-level SDMs.

Impact of spatial configuration of training data on the performance of Amazonian tree species distribution models

Remote sensing can provide useful explanatory variables for tree species distribution modeling, but only a few studies have explored this potential in Amazonia at local scales. Particularly for tropical forest management it would be useful to be able to predict the potential distribution of important tree taxa in areas where field data is as yet missing. Forest concessions produce valuable census data that cover large areas with high sampling effort and can be used as occurrence data in species distribution models (SDM). Nevertheless, these tree records are often spatially clumped and possibly only provide accurate predictions over areas close to where the training occurrence records are located. Here, we aim at investigating to what degree SDM performance and spatial predictions differ between models that have different spatial configurations of the occurrence data. For this, we divided the available occurrence data from a forest concession census in Peruvian Amazonia into different spatial configurations (narrow, elongated and compact), each of which contained approximately 20% of the full dataset. We then modelled the distributions of five tree taxa using Landsat data and elevation. More elongated configurations of the training data were more representative of the available environmental space, and also produced more robust SDMs. Average model performance (expressed as AUC) was 5% higher and variation in model performance 50% lower when elongated rather than compact configurations of training area were used. This confirms that covering only a small fraction of the environmental variability in the area of interest may lead to misleading SDM predictions, which needs to be taken into account when forest management decisions are based on SDMs.