Modelling Species Distributions Research Articles

Prediction of the Global Distribution of Arhopalus rusticus under Future Climate Change Scenarios of the CMIP6

Arhopalus rusticus is a significant forestry pest known for its destructive impact on various host plants. This species, commonly found in coniferous forests across the Northern Hemisphere, has successfully spread to regions like New Zealand, Australia, and South America. This research is based on the known distribution sites of A. rusticus. Projections are made for the potential global distribution of A. rusticus under historical climatic conditions (1970–2000) and future climatic conditions (2081–2100) for the four forcing scenarios of the Coupled Model International Comparison Program 6 (CMIP6). The aim was to analyze the effects of climate change on the distribution range of this pest and its invasion trend in the southern hemisphere, and to support relevant departments in enhancing the effectiveness of forestry pest control strategies. The study utilized the Biomod2 software package in R to compare six models: generalized linear models (GLMs), generalized additive models (GAMs), multivariate adaptive regression splines (MARSs), artificial neural networks (ANNs), classification and regression trees (CTAs), and random forests (RFs) for modeling species distributions. The optimal model was selected based on evaluation indexes such as AUC and TSS. Projections of A. rusticus distribution under historical and future climate scenarios were created. The prediction results were visualized using ArcGIS software (version 10.2) to classify fitness levels and calculate distribution areas. Based on evaluation metrics, random forests (RFs) demonstrated the highest average assessment index scores, indicating high prediction accuracy (AUC = 0.99, TSS = 0.91, Kappa = 0.93). Model predictions revealed that, under historical climatic conditions, A. rusticus was predominantly found in northern Europe, eastern Asia, eastern and southwestern coastal regions of North America, and there were also highly suitable regions in parts of the southern hemisphere, including central and southwestern Argentina, southern Australia, New Zealand, and South Africa. Among these models, each of the CMIP6’s different climate prediction scenarios had a significant impact on the predicted distribution of A. rusticus. The SSP126 scenario depicted the broadest range of suitability, while the SSP585 scenario presented the narrowest and, overall, the extent of highly suitable regions was contracting. Multi-model predictions suggested that the potential distribution area of A. rusticus during the period of 2081–2100 would likely expand compared to that of 1970–2000, ranging from an increase of 1.13% (SSP126) up to 6.61% (SSP585), positively correlating with the level of radiative forcing. Notably, the most substantial growth was observed in potentially low-suitability region, escalating from 1.17% (SSP126) to 5.55% (SSP585). The distribution of A. rusticus shows decreasing trends from coastal areas to inland areas and from high to low level suitability of regions, and further expansion into the southern hemisphere under future climate conditions. Therefore, quarantine efforts at ports of entry should be strengthened in areas that are not currently infested but are at risk of invasion, and precise preventive measures should be strengthened in areas that are at risk of further expansion under future climatic conditions to prevent its spread to inland areas.

Accurately predicting rare and poorly detectable species habitat for spatial protection

Abstract With the loss of biodiversity worldwide, understanding species distribution is essential for species management, but modelling the distribution of rare and poorly detectable species can be challenging because of data gaps and observer biases. Over‐ or under‐predictions are frequent, leading to uncertainty in spatial management measures, particularly for highly mobile and data‐poor species. Here we developed a ‘Combined Model for Accurate Prediction’ to accurately predict the distribution of data‐poor and rare species. This modelling framework aims to improve the accuracy of both predicted ‘core’ and ‘unsuitable’ habitats, to help support managers with spatial protection measures. We tested the combined modelling approach on 11 data‐poor and rare diadromous fish during their at‐sea life history phase and used the combined model to analyse the adequacy of existing marine protected areas (MPAs) for these fish. The combined modelling approach modelled both ‘core’ and ‘unsuitable’ habitats with high accuracy. Of the seven diadromous species modelled, most MPAs designated to protect diadromous fish are outside their core habitats. Furthermore, when their core habitat was within an MPA, only 50% of this area was designated to protect them. These results highlight inadequate protection of the existing networks of MPAs for protected and threatened species. Synthesis and applications. Being able to accurately model species distribution is critical to reliable and transparent biodiversity and conservation assessments. By modelling accurate ‘core’ and ‘unsuitable’ habitats with models that minimise omission and commission rates respectively, conservation measures could be targeted in specific spatial areas that maximise the protection of rare and poorly detected species. This method therefore helps minimise impacts on stakeholders, while providing managers with increased confidence in the model predictions.

A novel hybrid model for species distribution prediction using neural networks and Grey Wolf Optimizer algorithm

Neural networks are frequently employed to model species distribution through backpropagation methods, known as backpropagation neural networks (BPNN). However, the complex structure of BPNN introduces parameter settings challenges, such as the determination of connection weights, which can affect the accuracy of model simulation. In this paper, we integrated the Grey Wolf Optimizer (GWO) algorithm, renowned for its excellent global search capacity and rapid convergence, to enhance the performance of BPNN. Then we obtained a novel hybrid algorithm, the Grey Wolf Optimizer algorithm optimized backpropagation neural networks algorithm (GNNA), designed for predicting species’ potential distribution. We also compared the GNNA with four prevalent species distribution models (SDMs), namely the generalized boosting model (GBM), generalized linear model (GLM), maximum entropy (MaxEnt), and random forest (RF). These models were evaluated using three evaluation metrics: the area under the receiver operating characteristic curve, Cohen’s kappa, and the true skill statistic, across 23 varied species. Additionally, we examined the predictive accuracy concerning spatial distribution. The results showed that the predictive performance of GNNA was significantly improved compared to BPNN, was significantly better than that of GLM and GBM, and was even comparable to that of MaxEnt and RF in predicting species distributions with small sample sizes. Furthermore, the GNNA demonstrates exceptional powers in forecasting the potential non-native distribution of invasive plant species.

Climatic variables are more effective on the spatial distribution of oak forests than land use change across their historical range.

The current research is conducted to model the effect of climate change and land use change (LUC) on the geographical distribution of Quercus brantii Lindl. (QB) forests across their historical range. Forecasting was done based on six general circulation models under RCP 2.6 and RCP 8.5 future climate change scenarios for the future years 2050 and 2070. In order to model the species distribution, different modeling methods were used. The results indicated that, in general, climatic variables had a higher influence on the distribution of QB than land use-related attributes. The mean diurnal range (bio2), the precipitation seasonality (bio15), and the mean temperature of the driest quarter (bio9) were the main predictors in the distribution of QB forests, while land use variables were less important in oak species distribution. The GBM, MaxEnt, and RF had higher accuracy and performance in modeling species distribution. The outputs also showed that in the current climate circumstances, 97,608.81 km2 of the studied area has high desirability for the presence of QB, and by 2070, under the pessimistic scenario, 96.29% of these habitats will be lost under the concomitant effect of LUC and climate change. By using the results of this research, it is possible to predict and identify the effective factors in changing the habitat of this oak species with more certainty. Based on the insights obtained from the results of such studies, the protection and restoration planning of the habitat of this key species, which supports diverse species, will be provided more efficiently.

Integrating data from different taxonomic resolutions to better estimate community alpha diversity

Integrated distribution models (IDMs), in which datasets with different properties are analysed together, are becoming widely used to model species distributions and abundance in space and time. To date, the IDM literature has focused on technical and statistical issues, such as the precision of parameter estimates and mitigation of biases arising from unstructured data sources. However, IDMs have an unrealised potential to estimate ecological properties that could not be properly derived from the source datasets if analysed separately. We present a model that estimates community alpha diversity metrics by integrating one species‐level dataset of presence–absence records with a co‐located dataset of group‐level counts (i.e. lacking information about species identity). We illustrate the ability of community IDMs to capture the true alpha diversity through simulation studies and apply the model to data from the UK Pollinator Monitoring Scheme, to describe spatial variation in the diversity of solitary bees, bumblebees and hoverflies. The simulation and case studies showed that the proposed IDM produced more precise estimates of the community diversity than the single models, and the analysis of the real dataset further showed that the alpha diversity estimates from the IDM were averages of the single models. Our findings also revealed that IDMs had a higher prediction accuracy for all the insect groups in most cases, with this performance linked to the information provided by a data source into the IDM.

Top ten hazards to avoid when modeling species distributions: a didactic guide of assumptions, problems, and recommendations

Species distribution models, also known as ecological niche models or habitat suitability models, have become commonplace for addressing fundamental and applied biodiversity questions. Although the field has progressed rapidly regarding theory and implementation, key assumptions are still frequently violated and recommendations inadvertently overlooked. This leads to poor models being published and used in real‐world applications. In a structured, didactic treatment, we summarize what in our view constitute the ten most problematic issues, or hazards, negatively affecting implementation of correlative approaches to species distribution modeling (specifically those that model suitability by comparing the environments of a species' occurrence records with those of a background or pseudoabsence sample). For each hazard, we state relevant assumptions, detail problems that arise when violating them, and convey straightforward existing recommendations. We also discuss five major outstanding questions of active current research. We hope this contribution will promote more rigorous implementation of these valuable models and stimulate further advancements.

MaxEnt machine learning model predicts high groundwater potential areas in a fractured volcanic aquifer system

Groundwater constitutes a vital resource for public water supply, and thus, it is imperative to recognize the areas of highest potential for increasing availability. The present study employs the MaxEnt model to discern the most favorable areas for locating high-yield wells in Caxias do Sul, Rio Grande do Sul, southern Brazil, where the Serra Geral Aquifer System, a fractured volcanic aquifer, emerges. This aquifer system is characterized by its heterogeneous, discontinuous, and highly anisotropic nature. A dataset comprising 83 wells with high flow rates (≥10 m³/h) was selected from the municipal registry of deep tubular wells, along with 14 factors that influence groundwater occurrence (specific capacity, transmissivity, altitude, slope, horizontal curvature, vertical curvature, relief dissection index, drainage density, distance to drainage, topographic wetness index, distance to lineament, lineament density, precipitation, and soil hydrological group). The model output was a Groundwater Potential Map, which stochastically expresses the probability of obtaining flow rates ≥10 m³/h. The map was validated through cross-validation, resulting in an average accuracy of 65.14%, and by the Receiver Operating Characteristic analysis, resulting in an Area Under the Curve value of 0.911, indicating satisfactory validation. While the MaxEnt model is widely used in ecology to model species distribution, its application in groundwater prediction remains limited, particularly in fractured aquifers associated with volcanic rocks. Apart from optimizing the use of groundwater resources, this study also enhances the understanding of natural phenomena in this type of aquifer.

Thinning occurrence points does not improve species distribution model performance

AbstractSpatial biases are an intrinsic feature of occurrence data used in species distribution models (SDMs). Thinning species occurrences, where records close in the geographic or environmental space are removed from the modeling procedure, is an approach often used to address these biases. However, thinning occurrence data can also negatively affect SDM performance, given that the benefits of removing spatial biases might be outweighed by the detrimental effects of data loss caused by this approach. We used real and virtual species to evaluate how spatial and environmental thinning affected different performance metrics of four SDM methods. The occurrence data of virtual species were sampled randomly, evenly spaced, and clustered in the geographic space to simulate different types of spatial biases, and several spatial and environmental thinning distances were used to thin the occurrence data. Null datasets were also generated for each thinning distance where we randomly removed the same number of occurrences by a thinning distance and compared the results of the thinned and null datasets. We found that spatially or environmentally thinned occurrence data is no better than randomly removing them, given that thinned datasets performed similarly to null datasets. Specifically, spatial and environmental thinning led to a general decrease in model performances across all SDM methods. These results were observed for real and virtual species, were positively associated with thinning distance, and were consistent across the different types of spatial biases. Our results suggest that thinning occurrence data usually fails to improve SDM performance and that the use of thinning approaches when modeling species distributions should be considered carefully.

High-resolution data are necessary to understand the effects of climate on plant population dynamics of a forest herb.

Climate is assumed to strongly influence species distribution and abundance. Although the performance of many organisms is influenced by the climate in their immediate proximity, the climate data used to model their distributions often have a coarse spatial resolution. This is problematic because the local climate experienced by individuals might deviate substantially from the regional average. This problem is likely to be particularly important for sessile organisms like plants and in environments where small-scale variation in climate is large. To quantify the effect of local temperature on vital rates and population growth rates, we used temperature values measured at the local scale (in situ logger measures) and integral projection models with demographic data from 37 populations of the forest herb Lathyrus vernus across a wide latitudinal gradient in Sweden. To assess how the spatial resolution of temperature data influences assessments of climate effects, we compared effects from models using local data with models using regionally aggregated temperature data at several spatial resolutions (≥1 km). Using local temperature data, we found that spring frost reduced the asymptotic population growth rate in the first of two annual transitions and influenced survival in both transitions. Only one of the four regional estimates showed a similar negative effect of spring frost on population growth rate. Our results for a perennial forest herb show that analyses using regionally aggregated data often fail to identify the effects of climate on population dynamics. This emphasizes the importance of using organism-relevant estimates of climate when examining effects on individual performance and population dynamics, as well as when modeling species distributions. For sessile organisms that experience the environment over small spatial scales, this will require climate data at high spatial resolutions.

Equivalence of citizen science and scientific data for modelling species distribution of birds from a tropical savanna

AbstractThe Wallacean deficit continues to be a challenge to species distribution modelling. Although some authors have suggested that data collected by citizen scientists can be relevant for a better understanding of biodiversity, to our knowledge, no work has quantitatively tested the equivalence between scientific and citizen science data. Here, we investigate the hypothesis that data collected by citizen scientists can be equivalent to data collected by professional scientists when generating species spatial distribution models. For 42 bird species in the Cerrado region we generated and compared species distribution models based on three data sources: (1) scientific data, (2) citizen science data and (3) sample size corrected citizen science data. To test our hypothesis, we compared the equivalence of these datasets. We rejected the hypothesis of equivalence for about one‐third (38%) of the evaluated species, revealing that, for most of the species considered, the models generated were equivalent irrespective of the data set used. The distances between centroids of the models that were equivalent were on average smaller than the distances between non‐equivalent models. Also, the direction of change in the models showed no pattern, with no trend towards more populated regions. Our results show that the use of data collected by citizen scientists can be an ally in filling the Wallacean deficit gap. In fact, the lack of use of this wide range of data collected by citizen scientists seems to be an unjustified caution. We indicate the potential of using citizen science data for modelling the distribution of species, mainly due to the large set of data collected, which is impracticable for scientists alone to collect. Conservation measures will be favoured by the union of professional and amateur data, aiming for a better understanding of species distribution and, consequently, biodiversity conservation.

Effects of sample size, data quality, and species response in environmental space on modeling species distributions

Effects of sample size, data quality, and species response in environmental space on modeling species distributions

No country for small birds: Potential positive association among medium‐sized, aggressive species in Australian bird communities

AbstractAimAustralian woodlands have been intensively cleared since European settlement and, in parallel, many species of birds inhabiting this habitat type have experienced a marked decline. Conversely, some species such as noisy miner Manorina melanocephala respond positively to habitat disturbance and due to their hyper‐aggressiveness can end up driving away specialized and fragmentation‐sensitive species. Recent studies have suggested that the negative impact of miners is exacerbated by means of synergistic interactions between these and other aggressive species, including nest predators. However, it is not clear if these positive associations arise through similar habitat requirements or due to potential mutual benefits (‘facilitation effects’), which should predominate in harsh environments as the stress‐gradient hypothesis (SGH) predicts.LocationEastern Australia.MethodsWe combined a multi‐season community N‐mixture model and joint species distribution models within the Hierarchical Modelling of Species Communities framework to examine species co‐occurrence patterns after accounting for imperfect detection and the influence of environmental variables. We then explored how the balance between positive and negative associations varied along an abiotic stress gradient and whether species‐to‐species associations were non‐randomly distributed with respect to species traits.ResultsNo significant associations were detected; we only found tentative associations (posterior probability 85%) in a low proportion of species pairs (~4%). Although most non‐random (tentative) associations were negative, these were weaker and less consistent across models than the positive ones. Five medium‐sized species of swooping birds including the noisy miner monopolized virtually all positive associations. Communities from low‐productivity environments tended to show a lower degree of overall (community‐level) competitiveness than those located in less stressful (more productive) environments, which supports the SGH. There was no significant relationship between species trait dissimilarity and species‐to‐species association's strength.Main ConclusionsOur results suggest that aggressive species like miners, kookaburras, lorikeets, magpies and butcherbirds can form potential synergies with each other that may intensify the direct negative influence of each species on small‐sized songbirds. Since these species thrive in anthropized landscapes and have drastically increased their numbers in some regions, their associative potential should be considered in conservation actions.

Predict first–assemble later versus assemble first–predict later: Revisiting the dilemma for functional biogeography

Abstract Predicting community‐level trait indices, such as community‐weighted mean or functional dispersion, in space or time, is crucial, as they are markers of ecosystem functions, proxies for ecosystem services, and are now integral part of conservation planning. On one hand, since members of a species share similar traits, a natural way is to model species distributions as a function of the environment first (i.e. with species distribution models), and then to reconstruct trait indices for any locality of interest (predict‐first). On the other hand, community‐level trait indices can be seen as the direct result of environmental filtering, and thus their distributions may directly be modelled in response to the environment (assemble‐first). Although relatively different, these two approaches have been used interchangeably in trait‐based ecology with unknown consequences on their usability. Here, using plant community (4463 plots) and trait data (LNC, leaf nitrogen content; PLH, plant height; SLA, specific leaf area, for >800 species) covering the French Alps, we compared the two approaches to predict community mean and functional dispersion indices, accounting or not for abundance (CM, community mean; CWM, community weighted mean; FDis, functional dispersion; uFDis, unweighted functional dispersion). We tested both interpolation versus extrapolation capabilities of the two approaches. To support the empirical findings, we also run the same comparative analysis on simulated community data. While both approaches produced similar and skillful CM/CWM predictions (R2 > 0.63 for CWM PLH) for interpolation, the assemble‐first outperformed the predict‐first approach for extrapolation to unobserved environmental conditions (R2 = 0.60 against 0.54 for CWM PLH). Functional dispersion was generally less well predicted, although the assemble‐first approach again provided better predictions for both interpolation (R2 = 0.31 against 0.27 for FDis PLH) and extrapolation (R2 = 0.28 against 0.18 for FDis PLH). The predict‐first approach systematically overestimated FDis. Results were confirmed by the simulation experiments. We conclude that the direct modelling of community‐level indices provides more robust spatial predictions over space and time and should thus be preferred. This approach also does not suffer from having generally too many species predicted by the predict‐first approach.

Advancing terrestrial biodiversity monitoring with satellite remote sensing in the context of the Kunming-Montreal global biodiversity framework

Satellite remote sensing (SRS) provides huge potential for tracking progress towards conservation targets and goals, but SRS products need to be tailored towards the requirements of ecological users and policymakers. In this viewpoint article, we propose to advance SRS products with a terrestrial biodiversity focus for tracking the goals and targets of the Kunming-Montreal global biodiversity framework (GBF). Of 371 GBF biodiversity indicators, we identified 58 unique indicators for tracking the state of terrestrial biodiversity, spanning 2 goals and 8 targets. Thirty-six shared enough information to analyse their underlying workflows and spatial information products. We used the concept of Essential Biodiversity Variables (EBV) to connect spatial information products to different dimensions of biodiversity (e.g. species populations, species traits, and ecosystem structure), and then counted EBV usage across GBF goals and targets. Combined with published scores on feasibility, accuracy, and immaturity of SRS products, we identified a priority list of terrestrial SRS products representing opportunities for scientific development in the next decade. From this list, we suggest two key directions for advancing SRS products and workflows in the GBF context using current instruments and technologies. First, existing terrestrial ecosystem distributions and live cover fraction SRS products (of above-ground biomass, ecosystem fragmentation, ecosystem structural variance, fraction of vegetation cover, plant area index profile, and land cover) need to be refined using a co-design approach to achieve harmonized ecosystem taxonomies, reference states and improved thematic detail. Second, new SRS products related to plant physiology and primary productivity (e.g. leaf area index, chlorophyll content & flux, foliar N/P/K content, and carbon cycle) need to be developed to better estimate plant functional traits, especially with deep learning techniques, radiative transfer models and multi-sensor frameworks. Advancements along these two routes could greatly improve the tracking of GBF target 2 (‘improve connectivity of priority terrestrial ecosystems), target 3 (‘ensure management of protected areas’), target 6 (‘control the introduction and impact of invasive alien species’), target 8 (‘minimize impact of climate change on biodiversity’), target 10 (‘increase sustainable productivity of agricultural and forested ecosystems’) and target 12 (‘increase public urban green/blue spaces’). Such improvements can have secondary benefits for other EBVs, e.g. as predictor variables for modelling species distributions and population abundances (i.e. data that are required in several GBF indicators). We hope that our viewpoint stimulates the advancement of biodiversity monitoring from space and a stronger collaboration among ecologists, SRS scientists and policy experts.

Planning from scratch: A new modelling approach for designing protected areas in remote, data‐poor regions

Abstract Human pressure on ecosystems has strongly increased over the last decades and now impacts even the most remote regions. To help mitigate these impacts, it is crucial to designate protected areas in regions that retain a high level of ecological integrity. However, ecological data remain scarce for many such areas, making the systematic design of new protected zones challenging. Following a request from local managers, we developed an original methodological approach to help design new zoning for a pre‐existing protected area in a remote, data‐poor Sahelian wetland of southern Chad, a vast area rich in biodiversity and exploited by diverse human activities. The method involved first collecting extensive aerial survey data (6252 records) on birds and mammals and then analysing this through a combination of distance sampling and density surface modelling. The biodiversity data, combined with ecological predictors, helped model species distribution layers that were then incorporated with socio‐economic constraints into the systematic conservation planning tool Marxan. This approach produced an array of protected zoning options that met three levels of conservation objectives set by experts, corresponding to proportions of individuals from given species to protect in the proposed protected area. Frequent exchanges with local managers allowed the analyses to be refined, resulting in seven potential scenarios to be considered for conservation purposes. Synthesis and applications. In a context of high data scarcity, lack of access and short‐term conservation objectives, this combined approach that optimizes newly obtained data via a suite of modelling tools can facilitate identifying and protecting natural areas in regions most in need of urgent conservation policy.

Performance of inhomogeneous Poisson point process models under different scenarios of uncertainty in species presence-only data

Haphazard and opportunistic species occurrence (PO) data are widely used in species distribution models (SDMs) instead of high-quality species data gathered using appropriate and structured sampling methods, which is expensive and often spatially limited. Despite their widespread use in ecology, PO data are prone to errors and uncertainties, such as imperfect detectability, positional imprecision, and spatial niche truncation, which make their use analytically challenging for effective and adaptive biodiversity management and conservation. Using simulated data, this study investigates the effects of these uncertainties on the performance of spatial point process based presence-only and integrated SDMs. We investigated three SDMs in this study, one that ignores imperfect detectability: the presence-only model (PO model), and two that account for it: the thinned presence-only model (THINPO model) and the integrated model (PBPC model). The ability of these SDMs to produce accurate maximum likelihood estimates of intensity model coefficients and reliable predictions of species distributions under different data quality scenarios was investigated. The results show that SDMs that account for imperfect detectability (THINPO or PBPC models) are not applicable in situations of high detectability. In this situation, the PO model produces the most accurate maximum likelihood estimates of the models’ coefficients ({hat{beta }}_k), and consequently the most accurate predictions of species distributions ({hat{lambda }}(s)). The effects of positional uncertainty and spatial niche truncation on this SDM output are minimal. However, in situations of low detectability, it is preferable to use the PBPC model. Positional uncertainty and spatial niche truncation have negligible effects on the output of this SDM, except when positionally uncertain PO data are analyzed along with truncated PC data. These minimal effects of spatial niche truncation on SDM outputs demonstrate the transferability of SDMs. However, the effects of all these uncertainties may depend on the characteristics of the species. Prior to modeling species distributions, a multivariate environmental similarity surface analysis should be performed to test the similarity between data from the restricted region to be used for model calibration and data from the entire range. If this analysis reveals dissimilarities, larger spatial and ecological scales should be considered to address the issue of spatial niche truncation. Further efforts could address the effects of species characteristics on SDMs performance and assess the effects of species-specific uncertainties.

Effect of study area extent on the potential distribution of Species: A case study with models for Raoiella indica Hirst (Acari: Tenuipalpidae)

Ecological niche models are used to quantify the relationships between known occurrence records of a given species and environmental variables at these locations. Maxent is among the most widely used algorithms for modeling species distribution and has demonstrated better performance compared to other methods. However, the extent of the study area is a critical issue in the development of presence-only species distribution models because it encompasses the region used to extract the background points employed to characterize the environments accessible to the species. Thus, this study evaluated the effect of the extension of the study area on the species distribution modeling with the Maxent algorithm and occurrence data from the invasive species Raoiella indica Hirst (Acari: Tenuipalpidae). The increase in the study area extent inflated most of the threshold-dependent and -independent metrics used to assess model performance. The selection of the study area also affected the predicted suitable areas for the species (its potential distribution). The analysis shows that models developed with smaller study areas resulted in model overfitting and an increase in false-negative predictions. The extent of the area used during model training has a strong influence on the model outputs, with significant consequences for predicting the potential distribution of invasive species and thus for the areas under risk of invasion.

The unmarked R package: Twelve years of advances in occurrence and abundance modelling in ecology

AbstractSpecies distribution models (SDMs) are widely applied to understand the processes governing spatial and temporal variation in species abundance and distribution but often do not account for measurement errors such as false negatives and false positives.We describeunmarked, a package for the freely available and open‐source R software that provides a complete workflow for modelling species distribution and abundance while explicitly accounting for measurement errors. Here we focus on recent advances inunmarkedfunctionality to support multi‐species, multi‐state, and multi‐season data, as well as support for fitting models with random effects.For illustration, we present an analysis of Acadian FlycatcherEmpidonax virescensabundance on Roanoke River National Wildlife Refuge, North Carolina, USA, over 18 years. We found that Acadian Flycatcher abundance was initially greater in hardwood plantation habitat relative to bottomland hardwood forest along river levees but that abundance declined over time in both habitats.We plan forunmarkeddevelopment to keep pace with advances in hierarchical modelling in ecology, including better handling of continuous‐time data from camera trap and automated recording units and integrated models for multiple data streams.

Effectiveness of land use and disturbance measures, compared to climate, soil, and topography, for modeling relative abundances of tree species in the eastern United States

Environmental factors control species distributions and abundances, but effectiveness of land use and disturbance variables for modeling species generally is unknown compared to climate, soil, and topography variables. Therefore, I used predictor variables from categories of 1) land use and disturbance, 2) climate, and 3) soil, topography, and wind speed to model the relative abundances (i.e., percentage of all trees) of 65 common tree species in the eastern United States, with a contrast to presence-absence models of species distributions. First, I modeled variables within each category to identify the five most important variables. Then, I combined variables from each category to isolate most important variables, based on five model combinations of input variables from each category, ranging from one (i.e., three total) to five (i.e., 15 total) variables. From the five models of combined categories for each tree species, I identified the model with the greatest R2 value. Overall, climate variables were most important for tree species models with one and two input variables from each category, but land use and disturbance variables were most important for models with three to five input variables from each category. Although a range of R2 values occurred by species and number of input model variables, 32 species had best models with greatest R2 values of 0.50 to 0.81. For all best species models, the most important variables were temperature of the warmest quarter, historical fire return interval for all fires, agricultural area during years 1850 to 1997, and precipitation of the driest month. Current land cover classes, which are accessible and the most commonly modeled land use variables, were not important for modeling tree species abundances or distributions. Climate variables were most important for modeling species distributions. Results support the concept that while climate sets soft boundaries on distributions, relative abundances within distributions are affected by other filters. Future modeling may establish other important land use and disturbance variables, or refinements within the important variables of historical fire return interval and agricultural area over time, advancing integration of both land use and climate variables into studies.

Predicting the potential distribution of wheatear birds using stacked generalization-based ensembles

Habitat suitability models, usually referred to as species distribution models (SDMs), are widely applied in ecology for many purposes, including species conservation, habitat discovery, and gain evolutionary insights by estimating the distribution of species. Machine learning algorithms as well as statistical models have been recently used to predict the distribution of species. However, they seemed to have some limitations due to the data and the models used. Therefore, this study proposes a novel approach for assessing habitat suitability based on ensemble learning techniques. Three heterogeneous ensembles were built using the stacked generalization method to model the distribution of four wheatear species (Oenanthe deserti, Oenanthe leucopyga, Oenanthe leucura, and Oenanthe oenanthe) located in Morocco. Initially, a set of base-learners were constructed by virtue of training for each specie's dataset six machine learning algorithms (Multi-Layer Perceptron (MLP), Support Vector Classifier (SVC), K-nearest neighbors (KNN), Decision Trees (DT), Gradient Boosting Classifier (GB), and Random Forest (RF)). Then, the predictions of these base learners were fed as training data to train three meta-learners (Logistic Regression (LR), SVC, and MLP). To evaluate and assess the performance of the proposed approaches, we used: (1) six performance criteria (accuracy, recall, precision, F1-score, AUC, and TSS), (2) Borda Count (BC) ranking method based on multiple criteria to rank the best-performing models, and (3) Scott Knott (SK) test to statistically compare the performance of the presented models. The results based on the six-evaluation metrics showed that stacked ensembles outperformed their singles in all species datasets, and the stacked model with SVC as a meta-learner outperformed the other two ensembles. The results showed the potential of using ensemble learning techniques to model species distribution and recommend the use of the stacked generalization technique as a combination strategy since it gave better results compared to single models in four wheatear species datasets. Moreover, to assess the impact of future climate changes on the distribution of the four wheatear species, the best-performing distribution model was selected and projected into the current and future climatic conditions. The distributions of the Moroccan wheatear birds were found to be slightly affected by future climate changes.