Multiple Species Distribution Models Research Articles

Ensemble evaluation of potential distribution of Procambarus clarkii using multiple species distribution models.

Procambarus clarkii is a notorious invasive species that has led to ecological concerns owing to its high viability and rapid reproduction. South Korea, a country exposed to a high risk of introduction of invasive species due to active international trade, has suffered from recent massive invasions by invasive species, necessitating the evaluation of potential areas requiring intensive monitoring. In this study, we developed two different types of species distribution models, CLIMEX and random forest, for P. clarkii using occurrence records from the United States. The potential distribution in the United States was predicted along coastal lines and inland regions located below 40°N latitude The model was then applied to evaluate the potential distribution in South Korea, and an ensemble map was constructed to identify the most vulnerable domestic regions. According to both models, the domestic potential distribution was highest in most areas located at low altitudes. In the ensemble model, most of the low-altitude western regions, the eastern coast, and some southern inland regions were predicted to be suitable for the distribution of P. clarkii, and a similar distribution pattern was predicted when the model was projected into the future climate. Through this study, it is possible to secure basic data that can be used for the early monitoring of the introduction and subsequent distribution of P. clarkii.

Application of integrated species distribution models in parasitic diseases prevention and control: a review

Species distribution model, a mathematical model theoretically based on the temporal and spatial locations of species in the ecosystem and their associations with other species, is useful to display the current species distribution and predict the future species distribution, which has been widely applied in ecology, biogeography and spatial transmission of infectious diseases. Integrated species distribution model is a comprehensive, summative and reliable combination model that simultaneously uses multiple species distribution models or integrates multiple data sources for modeling. This review describes the theory and classification of species distribution models, and summarizes the creation and verification of integrated species distribution models and their applications in parasitic disease prevention and control. During the practical applications of integrated species distribution models, the study scale, niche matching, big data utilization, combination of multidisciplinary methods and knowledge background and ecological implications should be emphasized.

Optimising restoration and rehabilitation using environmental and spatial drivers of plant assemblages

The extent, condition and connectedness of ecosystems has been significantly impacted by human activities globally, leading to a widespread appeal of and demand for ecological restoration. Maximising the ecological outcomes and cost effectiveness of restoration requires that plans be optimised by indexing the distribution, abundance and condition of habitat-forming species, and that target ecosystem conditions be developed in concert with restoration site attributes. This type of quantitative approach for creating malleable ecological targets which are informed by local environmental conditions is, however, uncommon. In this study, we surveyed the composition and coverage of understorey plants, and the composition and size of trees in coastal dunes at five transects at each of 20 sites (for n = 100) on the Sunshine Coast, central eastern Australia, and use this information to prioritise restoration sites and create optimised planting regimes for degraded sites. Each of the identified indicator species (six understorey and four tree species) had unique preferred conditions and was affected by multiple environmental and spatial variables at varying spatial scales, with most species affected by urbanisation. Species distribution models (SDMs) were used to identify target/reference ecosystems and optimal planting mixes for potential restoration sites given the site’s environmental attributes. Our approach of integrating data on the distribution, abundance and condition of habitat-forming species into multiple SDMs can be used to optimise planting regimes at restoration sites and provides a framework for setting dynamic restoration targets across landscapes.

Predicting Potential Habitat Changes of Two Invasive Alien Fish Species with Climate Change at a Regional Scale

Developing national-level policies related to climate change induced expansions of invasive species requires predictive modelling at a regional scale level. This study aimed to predict future changes in the habitat distributions of two major invasive alien fish species, Micropterus salmoides and Lepomis macrochirus, in South Korea. An ensemble system with multiple species distribution models was used for the prediction, and gridded water portion data from the linear-structure information on river channels inputted as habitat characteristics of freshwater ecosystem into the models. Bioclimatic variables at 20-year intervals from 2001 to 2100 were generated from predicted temperature and precipitation data under the representative concentration pathway 4.5 and 8.5 scenarios. The overall distribution probabilities of the potential habitats increased with time in both climate change scenarios, and the potential habitats were predicted to expand to upstream areas. Combined with regional ecological value information, such as biodiversity in freshwater ecosystems, these results can be an important basis for deriving regional priority information for managing alien species in climate change. Additionally, the modelling approach is highly applicable to various national-level policies for ecosystem conservation since it is not greatly restricted by spatial scales.

Using a multi-model ensemble approach to determine biodiversity hotspots with limited occurrence data in understudied areas: An example using freshwater mussels inMéxico.

Species distribution models (SDMs) are an increasingly important tool for conservation particularly for difficult‐to‐study locations and with understudied fauna. Our aims were to (1) use SDMs and ensemble SDMs to predict the distribution of freshwater mussels in the Pánuco River Basin in Central México; (2) determine habitat factors shaping freshwater mussel occurrence; and (3) use predicted occupancy across a range of taxa to identify freshwater mussel biodiversity hotspots to guide conservation and management. In the Pánuco River Basin, we modeled the distributions of 11 freshwater mussel species using an ensemble approach, wherein multiple SDM methodologies were combined to create a single ensemble map of predicted occupancy. A total of 621 species‐specific observations at 87 sites were used to create species‐specific ensembles. These predictive species ensembles were then combined to create local diversity hotspot maps. Precipitation during the warmest quarter, elevation, and mean temperature were consistently the most important discriminatory environmental variables among species, whereas land use had limited influence across all taxa. To the best of our knowledge, our study is the first freshwater mussel‐focused research to use an ensemble approach to determine species distribution and predict biodiversity hotspots. Our study can be used to guide not only current conservation efforts but also prioritize areas for future conservation and study.

A winner or a loser in climate change? Modelling the past, current, and future potential distributions of a rare charophyte species

Due to human–caused eutrophication and rapid environmental changes, charophytes have become an increasingly rare component of biosphere. One of the rarest charophyte worldwide is Lychnothamnus barbatus , the only extant representative of the genus Lychnothamnus . The species occurs in Europe, Asia, Australia and North America, but the number of locations is scarce and the geographical distribution is uncertain. In this study, multiple species distribution models (SDMs: Bioclim, MaxEnt, random forest (RF) and support vector machines (SVMs)) were used to reveal the potential undiscovered locations and ranges of L. barbatus . We aimed to (1) indicate the most reliable SDM for L. barbatus ; (2) determine the most important bioclimatic factor(s) affecting this species distribution; and (3) estimate the current, past (ca. 130 ka) and future (up to 2100) habitat suitability of L. barbatus in Europe and North America. Future climate scenarios were calculated using four global climate models (GCMs): CanESM5, CNRM–CM6–1, MIROC 6, BCC-CSM2-MR and model ensemble. We documented that a SVM model, which is relatively rarely used in SDMs, performed better than commonly used MaxEnt and RF models. The most important predictors for the SVM (i.e., temperature in wettest, warmest and driest quarters; precipitation in warmest and wettest quarters) were partly different than those for MaxEnt (temperature in wettest quarter and annual temperature range). The SVM used predictors more efficiently than did MaxEnt to achieve good performance. Thus, we recommend further exploring the potential of SVMs in SDMs. Furthermore, we tested different GCMs and demonstrated that the predicted future range of Lychnothamnus is heavily dependent on the model type. Nevertheless, all the models indicated expansion of L. barbatus to other areas in Europe when the pessimistic scenario occurred. It is important for biological conservation of currently rare species—more effort could be put to conserve other, warming–vulnerable species. • As rare hydrobionts, charophytes are vulnerable to climate warming effects. • Past, present and future distribution of Lychnothamnus barbatus was modeled. • Multiple species distribution models, global climate models and scenarios were used. • This study evidenced that the rare L. barbatus may expand its range in the future. • A method to detect warming-vulnerable species among currently rare species proposed.

Cetacean distribution models based on visual and passive acoustic data

Distribution models are needed to understand spatiotemporal patterns in cetacean occurrence and to mitigate anthropogenic impacts. Shipboard line-transect visual surveys are the standard method for estimating abundance and describing the distributions of cetacean populations. Ship-board surveys provide high spatial resolution but lack temporal resolution and seasonal coverage. Stationary passive acoustic monitoring (PAM) employs acoustic sensors to sample point locations nearly continuously, providing high temporal resolution in local habitats across days, seasons and years. To evaluate whether cross-platform data synthesis can improve distribution predictions, models were developed for Cuvier’s beaked whales, sperm whales, and Risso’s dolphins in the oceanic Gulf of Mexico using two different methods: generalized additive models and neural networks. Neural networks were able to learn unspecified interactions between drivers. Models that incorporated PAM datasets out-performed models trained on visual data alone, and joint models performed best in two out of three cases. The modeling results suggest that, when taken together, multiple species distribution models using a variety of data types may support conservation and management of Gulf of Mexico cetacean populations by improving the understanding of temporal and spatial species distribution trends.

Predicting the distributions of regional endemic dragonflies using a combined model approach

Abstract Climate warming is predicted to have large effects on insects, yet several data shortfalls, including distributional information, impede effective conservation strategies. Knowledge of species distributions is a critical component for assessing conservation need but is often lacking for endemic or rare taxa, especially invertebrates. One approach to better inform this gap is by using species distribution modelling (SDM) to predict suitable habitat and guide field surveys. Here, we combine the predictions of two machine learning algorithms, maximum entropy and Random Forest, to estimate the current and future distributions of two endemic dragonflies of the Ozark‐Ouachita Interior Highlands region in the southcentral United States. Current suitable areas predicted by both algorithms largely overlapped for each species, but different environmental variables were most important for predicting their distributions. Field validation of these models resulted in new detections for both species showing their utility in guiding subsequent field surveys. Future projections under two climate change scenarios support maintaining current suitable areas as these are predicted to be strongholds for these species. Our results suggest that combining outputs of multiple species distribution models is a useful tool for better informing the distributions of geographically limited or rare species.

Effects of environmental factors and heavy metals on the vertical distribution of bryophytes in a sinkhole environment.

A sinkhole ecosystem, as a refuge for plant diversity, has been subjected to intensive exploitation, leading to ecosystem destruction of sinkholes in China. Understanding the responses of bryophyte distribution to destruction of the sinkhole environment are crucial to implementing protection measures for bryophyte diversity. Haolong sinkhole in Guangxi Zhuang Autonomous Region of China, the third largest sinkhole in the world, was selected as the study area. The Wilson Shmida index was used to analyse bryophyte species diversity; a Generalized Linear Model (GLM) was used to reveal species vertical distribution of bryophytes, the Single and Multiple Species Distribution Models (SSDM, MSDM) were used for analysis of the relationship between bryophyte species distribution, environmental factors and heavy metals. A total of 183 species from 74 genera in 36 families of bryophytes were collected from Haolong sinkhole, of which 26 species are endemic to China. Bryophyte species diversity was ranked in the order: agricultural section<forest section<grassland. In the vertical direction, bryophyte distribution was divided into point, disjunctive and continuous distributions using the GLM. The SSMA and MSDM indicated that bryophyte species of each of these three distributions can be divided into a temperature-slope zone, light-depth-pH-humidity zone, Pb (B)-Hg (B) zone and mixed heavy metals zone according to the effect of environmental factors and heavy metals such as As. Environmental factors or heavy metals, such as As, in Haolong sinkhole effectively cooperate in bryophyte distribution. An effective way to protect bryophyte diversity, in particular species endemic to China in the sinkhole environment, is through education and involvement of the local villagers to minimize further damage to the sinkhole environment.

Modelling the spatial distribution of three Portunidae crabs in Haizhou Bay, China.

Crab species are economically and ecologically important in coastal ecosystems, and their spatial distributions are pivotal for conservation and fisheries management. This study was focused on modelling the spatial distributions of three Portunidae crabs (Charybdis bimaculata, Charybdis japonica, and Portunus trituberculatus) in Haizhou Bay, China. We applied three analytical approaches (Generalized additive model (GAM), random forest (RF), and artificial neural network (ANN)) to spring and fall bottom trawl survey data (2011, 2013–2016) to develop and compare species distribution models (SDMs). Model predictability was evaluated using cross-validation based on the observed species distribution. Results showed that sea bottom temperature (SBT), sea bottom salinity (SBS), and sediment type were the most important factors affecting crab distributions. The relative importance of candidate variables was not consistent among species, season, or model. In general, we found ANNs to have less stability than both RFs and GAMs. GAMs overall yielded the least complex response curve structure. C. japonica was more pronounced in southwestern portion of Haizhou Bay, and C. bimaculata tended to stay in offshore areas. P. trituberculatus was the least region-specific and exhibited substantial annual variations in abundance. The comparison of multiple SDMs was informative to understand species responses to environmental factors and predict species distributions. This study contributes to better understanding the environmental niches of crabs and demonstrates best practices for the application of SDMs for management and conservation planning.

A suite of macrophyte species distribution models for investigating hydrology-driven spatial changes in a large flood-pulsed tropical wetland

ABSTRACTSpatial ecological models of wetland systems are important tools for evaluating system response to changes in hydrological inputs, arising from anthropogenic manipulation or climate change. We present an approach using multiple species distribution models (SDMs) for floodplain macrophytes, based on empirical relationships with flood duration. 27 indicator species were identified for four floodplain classes in a seasonally pulsed distributary of the Okavango Delta, based on data from a field survey of 30 floodplain sites. Spatial hydrological history was constructed using a combination of Landsat and MODIS imagery, and ground-truthing. Relationships between duration and occurrence were used to construct general linear probability models (GLMs) of species along the hydroperiod gradient. The GLM parameters were then used to construct SDMs driven by a spatial hydrological model, which generates annual assessments of inundation duration in months in each 1km2 cell across the Delta. The modelled distribution fits were mostly robust (18 GLMs had p < 0.05). Tests of model performance against a validation set of five floodplain sites showed an average of 87% correct predictions of presence. The strong performance of the models and the ability to produce map-based and quantitative output indicate good potential for use in assessment of basin development scenarios.

Assessing uncertainty and performance of ensemble conservation planning strategies

Systematic conservation initiatives attempt to cater to the needs of many species via the integration of multiple species distribution models (SDMs), or via the integration of Systematic Conservation Planning (SCP) software, such as Zonation. Unfortunately, due to limited data and knowledge, it is often difficult to select the most suitable model for specific species, let alone an appropriate ensemble modeling method for multiple species. In general, model selection criteria are based on either model performance or consensus. The former integrates the highest-performing SDM for all focal species, whereas the latter integrates multiple SDM outputs based on consensus. While higher-performing ensemble models presumably identify high-quality habitats better, many have argued that high consensus ensemble models have less uncertainty originating from sporadic model variability. This study develops and validates seven ensemble-modeling strategies for integrating outputs of the systematic conservation tool Zonation. First, we considered the distributions of 11 bird species via 100 runs of five SDMs across Taiwan. Second, we evaluated the local and global uncertainty of all five models. Third, we used Zonation to obtain conservation priorities. We then used Principal Component Analysis (PCA) to quantify different sources of uncertainty. Finally, we used independent third-party habitat data to validate each strategy. On average, the ‘best model’ strategy (based on the highest AUC value) performed best. Based on our modeling exercise we present a comprehensive framework for conservation prioritization, validation and the quantification of uncertainty intrinsic to SDMs according to different conservation scenarios and goals.

Maps, models, and marine vulnerability: Assessing the community distribution of seabirds at-sea

Whether considering the cumulative impact of chronic, small-size oil discharges or accidents associated with marine traffic and offshore oil and gas development, seabirds face a variety of threats in the marine environment. Assessing the vulnerabilities of seabirds to maritime hazards requires an understanding of their species distribution, and a means for combining this information across groups. Using at-sea survey data gathered as part of a regional monitoring program, an efficient framework for integrating multi-species data was developed. Survey data was incorporated within a distance sampling framework to generate bias-corrected seabird densities for an area of over 730,000km2 size, which were used to construct multiple species distribution models (SDMs). The structural difficulties associated with sampling sparsely distributed individuals that also occur in large, localized concentrations led to the use of three modelling techniques potentially well suited for this type of data: negative binomial, “hurdle”, and random forest methods. Predicted abundances were combined to produce an ensemble forecast, which met or exceeded the accuracy of predictions from the individual models. Multi-species potential sensitivity maps were developed to identify core areas, confirming the general importance of physiographic features such as the shelf break and bathymetry. Distribution was also seasonally influenced, with spring and winter standing out as periods of peak importance. When combined with an oil pollution layer derived from aerial surveillance, vulnerability was highest in the vicinity of major ports (e.g., Halifax and Sydney, Nova Scotia). However, the vulnerability map also indicated lower but widespread levels of oiling risk throughout the shelf, presumably associated with persistently high levels of shipping traffic and ongoing petroleum extraction and exploration. Outside of this region, migratory connectivity is expected to expose seabirds to a wider network of hazards and further underscores the need for the coordinated and routine collection of marine hazard data alongside at-sea distributional data.

Where the bears roam in Majella National Park, Italy

The Brown bear (Ursus arctos Linnaeus, 1758) occupies contiguous areas in Eastern and Northern Europe. In Western Europe, the largest remnant populations occur in Cantabria, Spain and the Apennines, Italy. Under Italian law the bear and its occupied range are protected. The occupied range of the Apennine brown bear includes Majella National Park. However, information on the distribution within Majella NP is extrapolated and inconsistent, thus precluding evidence-based protected area and species management. To address this lack of information, bear presence records (1996–2010) were collated and corrected for observational bias. Multiple Species Distribution Models (SDMs) were created at 800m resolution to predict year-round and seasonal bear distribution. A hierarchical, stepwise maxent SDM approach was applied using climatic, terrain, vegetation, and anthropogenic predictors of bear distribution. Occupied ranges were identified by point density analysis of bear presence. Our climate-only SDMs predicted bear presence in areas with relatively low snowfall and temperate temperatures. Year-round bear distribution was also accurately predicted by using temperate-montane elevations and mesic, mesotrophic vegetation substrates, irrespective of vegetation. Ski-resorts were negative predictors of year-round bear occurrence. Bears were predicted in autumn and winter by beech forest, in spring by meadows and in summer by a variety of vegetation categories. The regional and our local models predicted bear throughout the south. However, our predicted and occupied range in the north includes the Orta valley and exclude alpine heights, contrary to the regional models. Only our summer bear range is similar to a regional SDM. We demonstrated that multiple maxent SDMs using a modest number of observations and a comprehensive set of environmental variables may generate essential distributional information for protected area and species management where full wildlife and food source censuses are lacking.

Predicting the Distribution of Commercially Important Invertebrate Stocks under Future Climate

The future management of commercially exploited species is challenging because techniques used to predict the future distribution of stocks under climate change are currently inadequate. We projected the future distribution and abundance of two commercially harvested abalone species (blacklip abalone, Haliotis rubra and greenlip abalone, H. laevigata) inhabiting coastal South Australia, using multiple species distribution models (SDM) and for decadal time slices through to 2100. Projections are based on two contrasting global greenhouse gas emissions scenarios. The SDMs identified August (winter) Sea Surface Temperature (SST) as the best descriptor of abundance and forecast that warming of winter temperatures under both scenarios may be beneficial to both species by allowing increased abundance and expansion into previously uninhabited coasts. This range expansion is unlikely to be realised, however, as projected warming of March SST is projected to exceed temperatures which cause up to 10-fold increases in juvenile mortality. By linking fine-resolution forecasts of sea surface temperature under different climate change scenarios to SDMs and physiological experiments, we provide a practical first approximation of the potential impact of climate-induced change on two species of marine invertebrates in the same fishery.

Geographic shifts in climatically suitable areas and loss of genetic variability in Dipteryx alata (“Baru” Tree; Fabaceae)

Many species are expected to suffer strong shifts in their geographic ranges due to climate changes in the next 50 years, with severe consequences for biodiversity patterns and population structure. We used here an ensemble forecast approach for obtaining species' range in which multiple species distribution models and climatic models were combined to model loss of genetic variability in Baru, Dipteryx alata (Fabaceae), an economically important Neotropical tree native to the Cerrado of Brazil. We estimated a series of genetic parameters (number of alleles per locus, expected heterozygosity under Hardy-Weinberg equilibrium and mutation-drift equilibrium) for this species based on eight microsatellite loci. We then recalculated these parameters assuming that local populations in areas of low future environmental suitability will go extinct. All genetic parameters remained approximately constant up to a 50% threshold of climatic suitability in the future; after this critical threshold there is an abrupt reduction in all parameters, although the magnitude of shift is only about 10% of current values, on average. Thus, despite the shifts in geographic range and climatically suitable areas towards southeastern Brazil, our analyses do not predict a strong loss of genetic diversity in D. alata because of the broad tolerance of this species, which ensures large future ranges, contrasting with other Cerrado species that have been analyzed in a similar manner.