Article
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Aim: To investigate how changing grid size can alter model predictions of the distribution of mesophotic taxa and how it affects different modelling methods. Location: Ningaloo Marine Park, Western Australia. Taxon: Benthic mesophotic taxa: corals, macroalgae and sponges. Methods: We determined the distributions of the major benthic taxonomic groups: corals, macroalgae and sponges, using a number of modelling techniques and an ensemble using the ‘sdm’ R package. A range of grid sizes were used (10, 50, 100 and 250 m) to identify how model predictions were altered. Models were evaluated using the area under the curve of a receiver operator characteristic plot (AUC) and the true skill statistic (TSS) using a spatially independent dataset. Results: Grid size had a large effect on model performance across the taxonomic groups. Model outputs were compared to null surfaces and 88.8% of models performed significantly better than null. Distribution of corals was best predicted using the finest grid size (10 m) regardless of modelling method, although a model ensemble produced the best results (AUC = 0.80, TSS = 0.52). Macroalgae and sponges were better predicted at coaster grids sizes (250 m). Again, ensembles performed well for both macroalgae (AUC = 0.83, TSS = 0.63) and sponges (AUC = 0.88, TSS = 0.66). Model ensembles maintained high accuracy across grid sizes and were consistently the best, or second‐best, performing method. Main conclusions: This study has shown how grid size should be considered when producing distribution models. Identifying the most relevant grid size and being aware of the influence it may have will provide more accurate predictions of the distributions of taxa. Ensemble methods maintained good performance across scenarios and thus provide a useful tool for conservation and management especially where single modelling methods showed high levels of variability.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... The only solution for regional studies, as in this case, is to assume no distribution shifts over time. Working at regional analyses may provide less detailed results (Bennie et al. 2014, Turner et al. 2019), but that does not invalidate the present results. Alternatively, citizen science data from online platforms may help (e.g. ...
Article
Full-text available
Climate change has strong effects on biodiversity. It causes shifts in the species' geographical range, towards other latitudes or elevations, while tracking its natural habitat. Here, the species maintains its realised niche stable (in the environmental space) and shifts its geographical range. Alternatively, a species might shift its realised niche through its fundamental niche, maintaining its geographical range stable, when tracking its natural habitat is not possible. Following this last hypothesis, we aimed to assess changes in species' environmental conditions based on their historical distribution. Within the environmental space, we compared the climatic realised niche of 300 species of several groups of flora and fauna between 1979 and 2013 in the Iberian Peninsula with ecospat R package, under the assumption of no distribution shift over time. We used the same species' distribution data in both periods. We used CHELSA yearly averaged time‐series data for 1979 and 2013. We analysed the differences in the climate between 1979 and 2013 with a Kruskal–Wallis test, by subtracting each pair of variables, and by subtracting the first three components of a PCA. For each species, we measured the realised niche similarity with Schoener's D index, the realised niche overlap with the stability, unfilling and expansion indices, and the distance between realised niche centroids. The climate between 1979 and 2013 changed in the Iberian Peninsula: climate variability reduces in 2013. Most of the species do not have equivalent realised niches. Although very few species had completely dissimilar niches, many species have shifted their realised niches to some degree. In only 35 years, it was already possible to observe how the realised niches of some species started to shift.
... The ensemble modeling was suggested as among the robust approaches in distribution modeling (Araújo and New 2007;Gómez et al. 2018;Hao et al. 2020;Marmion et al. 2009). The ensemble modeling approach has the potential to minimize the expected modeling result variability which may occur when using a single algorithm (Alfaro et al. 2018;Buisson et al. 2010;Turner et al. 2019). An ensemble model has been reported as the most efficient model in predicting different alien invasive species by Stohlgren et al. (2010) and Ng et al. (2018). ...
Article
Full-text available
Background Senna didymobotrya is a native African flowering shrub. It is suspected that climate change encourages the introduction and spread of invasive alien species. The possible invasion of S . didymobotrya across the continent is expected to increase in the future due to ongoing climate change. Nonetheless, there is still paucity of empirical evidence on the extent to which the changing climate contributes to the surge of the flowering shrub. This study, therefore, investigated the present and potential invasion of S . didymobotrya using the species distribution model under changing climate conditions. The two representative concentration pathways (RCP4.5 and RCP8.5) and eight bioclimatic variables and one topographic variable were used to simulate the current and future (2050s and 2070s) invasion of S . didymobotrya in Africa. The model performance was assessed using the area under the receiver operating characteristic curve (AUC) and true skill statistics (TSS). Results The results of the study showed that under the current climatic conditions, 18% of Africa is suitable for the establishment and invasion of S . didymobotrya . The most suitable hotspot for S . didymobotrya invasion is eastern Africa, followed by southern Africa. The predicted model showed that by 2050, 3.3% and 3.12% of the continent would be highly suitable areas for the invasion of the species under RCP4.5 and RCP8.5, respectively. In the 2070s, under RCP4.5 and RCP8.5, the highly suitable area would be 3.13% and 2.7%, respectively. In relation to the current suitability, the cumulative projected areas of the low and moderate suitability class under RCP4.5 and RCP8.5 will rise by the years 2050 and 2070. However, under both RCPs, the non-suitable area for S . didymobotrya invasion would gradually decrease. Conclusions From the findings, it can be concluded that the ecosystem’s vulnerability to S . didymobotrya invasion under future climatic conditions will proliferate significantly. Hence, to prevent the projected harm to biodiversity and ecosystem services, governments need to focus their future biodiversity management and policy directions on the means and strategies of minimizing the invasion and the distribution rate of S . didymobotrya across habitat types.
... ENMs are powerful statistical tools for analysing species niches and their changes over time and space [22]. However, very few studies have applied ENMs at very high spatial resolution [23][24][25][26][27][28] and even less have compared several species composing a community [10]. ...
Article
Full-text available
Species can occupy different realised niches when sharing the space with other congeneric species or when living in allopatry. Ecological niche models are powerful tools to analyse species niches and their changes over time and space. Analysing how species’ realised niches shift is paramount in ecology. Here, we examine the ecological realised niche of three species of wall lizards in six study areas: three areas where each species occurs alone; and three areas where they occur together in pairs. We compared the species’ realised niches and how they vary depending on species’ coexistence, by quantifying niche overlap between pairs of species or populations with the R package ecospat. For this, we considered three environmental variables (temperature, humidity, and wind speed) recorded at each lizard re-sighting location. Realised niches were very similar when comparing syntopic species occurring in the same study area. However, realised niches differed when comparing conspecific populations across areas. In each of the three areas of syntopy, the less abundant species shift its realised niche. Our study demonstrates that sympatry may shift species’ realised niche.
Article
Full-text available
Bamboo is the world's most widely exploited plant resource, with significant socio-economic and cultural values. In most parts of Africa, the population is in jeopardy due to the high pressure from human and natural forces. Of these, Oxytenanthera abyssinica (A. Richard) is among the threatened bamboo species. Furthermore, the effect of climate change on the distribution of bamboo has not yet been adequately studied. Therefore, this study aims to model and map the current and future distribution of O. abyssinica in Africa under four representative concentration pathways (RCPs), such as RCP2.6, RCP4.5, RCP6, and RCP8.5. The future projections were done for the years 2050 and 2070 using SDM ensemble approaches. To model the current and future distribution of O. abyssinica in Africa, 737 presence data were collected from various sources. For this study, a total of eight (8) temperature and precipitation-related variables were used as inputs to the Species Distribution Model (SDM). Finally, the model performance was assessed based on the area under the curve (AUC) and true skills statistics (TSS) measures of statistics. Our results showed an upsurge in the distribution of O. abyssinica across the study area for the low and moderate suitability classes for the climatic conditions considered in this study. However, a steady shrinkage in the habitat was found for the higher suitability classes. The model indicated climatic-related factors such as precipitation during the cold and warm quarters (57.8%), followed by mean temperature during the coldest quarter, isothermality (41.9%) and topographic factors such as elevation and slope (31.6%) were identified as the main limiting factors for the growth of O. abyssinica. Precipitation and temperature during the dry period, on the other hand, had the least impact on the growth of O. abyssinica. Except for RCP2.6, the majority of south-western African countries and the Sahel region remain the most climatically stable ecosystems for O. abyssinica growth under the three climatic scenarios RCP45, RCP6 and RCP8.5. Our results revealed a steady increase in the future suitable habitat for O. abyssinica all over the continent under the considered climatic scenarios. Therefore, to support the future restoration of dryland ecosystems, countries should scheme a restoration policy that allows the sustainable utilization of O. abyssinica tree species. The future policy direction for biodiversity conservation and management should encourage the use of O. abyssinica as a major plant species for improving the livelihoods of people living in dryland areas.
Thesis
Full-text available
The intertidal zone is affected by environmental change in both the marine and the terrestrial realms. All along Europe’s coast, a unique biogenic habitat buffers the shores’ environmental extremes for the high-diversity fauna it hosts: the reefs of the honeycomb worm Sabellaria alveolata. Despite being listed as a priority European Union habitat, the criteria for assessing candidate reefs have yet to be defined. As a prerequisite to defining the habitat’s ecological status, the aim of this thesis was to further our knowledge of S. alveolata’s ecological niche through an integrated, multi-scale approach. Its niche was explored through the identification and analysis of physiological condition biochemical indicators, which were linked to the outward physical appearance of the bioconstructions in which the individuals resided. Spatial and intraspecific variation in reproductive and biochemical traits were analysed in individuals sampled in ten sites between North Wales and central Portugal. This made it possible to test two divergent theories on the optimal position and environmental conditions for reproduction within a species range. Finally, by constructing a broad-scale occurrence records database which served as the response variable in a species global distribution model, predictive ‘range-long’ maps of present and future suitable habitats were generated for the whole north-east Atlantic coastline. By integrating ecophysiological and biogeographical approaches, this thesis has brought mechanistic insight to the broad-scale processes shaping the distribution of this key habitat-forming species.
Article
Full-text available
The Southern Ocean faces many challenges under global climate change. Species distribution models (SDMs) are considered a powerful tool for predicting the potential distributions of species when evaluating the effects of climate change. We employed a partially area-under-the-curve (AUC)-based ensemble model to predict the current habitat suitability and future habitat changes of N. coatsorum in the Southern Ocean. The SDMs developed included six different predictors that contribute to the N. coatsorum distribution: depth accounted for the highest contribution of 33.8%, closely followed by the sea surface temperature (30.3% contribution). The contributions of the sea surface salinity distance from land, sea ice thickness and current velocity were 12.7%, 12.3%, 8.1%, and 2.8%, respectively. N. coatsorum has a suitable distribution area of approximately 5.63 million km² under current conditions, comprising29.8% of the Southern Ocean area. A significant expected decrease(43.0%) at the species’ northward edges, especially on the southeast Antarctic coast, may occur in the future as suggested by comparing the current model range to the suitable habitat changes predicted in 2100 under the RCP8.5 scenario, resulting in the formation of three newly gained or remaining climate refugia for N. coatsorum: areas close to the Ronne and Filchner Ice Shelves in the Weddell Sea, the Ross Sea, and along the coast of northeastern Antarctica (including in the Cosmonaut Sea and Prydz Bay). Our ensemble approach for assessing the habitat features of N. coatsorum allows the impacts of climate change on a mesopelagic fish species in the Southern Ocean to be assessed.
Article
Full-text available
Despite a growing understanding of the importance of mesophotic ecosystems, they remain relatively unexplored globally, and particularly in the Indian Ocean. The composition of benthic communities of the Ningaloo Marine Park in deeper water (> 20 m) was determined using autonomous underwater vehicles (AUVs). Environmental variables collected by the AUVs as well as geomorphological variables derived from high-resolution multibeam bathymetry were used to explain the distribution of these communities. Imagery was classified for benthic biota, and non-hierarchical k-means clustering was used to identify community groups. Overall, 12 community groups were identified which fell into three broad categories: (1) coral-dominated, (2) CCA-dominated; and (3) macroalgae- and filter-feeder-dominated. Specific deeper water communities (> 50 m) were characterised by the presence of sponges or other non-photosynthetic filter feeders. Corals were confined to depths < 40 m, and cover was high (29–40%). Generalised additive models identified the importance of substrate (cover of sand, rock, and gravel/pebbles), geomorphological variables (slope and bathymetric position index) and temperature in the distribution of communities.
Article
Full-text available
Species distribution models have been used to predict the distribution of invasive species for conservation planning. Understanding spatial transferability of niche predictions is critical to promote species-habitat conservation and forecasting areas vulnerable to invasion. Grain size of predictor variables is an important factor affecting the accuracy and transferability of species distribution models. Choice of grain size is often dependent on the type of predictor variables used and the selection of predictors sometimes rely on data availability. This study employed the MAXENT species distribution model to investigate the effect of the grain size on model transferability for an invasive plant species. We modelled the distribution of Rhododendron ponticum in Wales, U.K. and tested model performance and transferability by varying grain size (50 m, 300 m, and 1 km). MAXENT-based models are sensitive to grain size and selection of variables. We found that over-reliance on the commonly used bioclimatic variables may lead to less accurate models as it often compromises the finer grain size of biophysical variables which may be more important determinants of species distribution at small spatial scales. Model accuracy is likely to increase with decreasing grain size. However, successful model transferability may require optimization of model grain size.
Article
Full-text available
Mesophotic coral ecosystems (MCEs) occur at depths beyond those typically associated with coral reefs. Significant logistical challenges associated with data collection in deep water have resulted in a limited understanding of the ecological relevance of these deeper coral ecosystems. We review the trends in this research, covering the geographic spread of MCE research, the focus of these studies, the methods used, how MCEs differ in terms of species diversity and begin to assess connectivity of coral populations. Clear locational biases were observed, with studies concentrated in a few discrete areas mainly around the Atlantic region. The focus of MCE studies has diversified in recent years and more detailed aspects of MCE ecology are now being investigated in particular areas of research. Advances in technology are also reflected in the current range of research, with a wider variety of methods now employed. However, large information gaps are present in entire regions and particularly in relation to the threats, impacts and subsequent management of MCEs. Analysis of species diversity shows that initial definitions based on depth alone may not be appropriate globally, while further taxonomic resolution may also be required to deduce the full biodiversity of major groups in certain regions. Genetic studies to date show species-specific results, although distinct deeper populations do appear to exist, which raises questions regarding the potential of MCEs to act as refugia.
Article
Full-text available
In the deep sea, biological data are often sparse; hence models capturing relationships between observed fauna and environmental variables (acquired via acoustic mapping techniques) are often used to produce full coverage species assemblage maps. Many statistical modelling techniques are being developed, but there remains a need to determine the most appropriate mapping techniques. Predictive habitat modelling approaches (redundancy analysis, maximum entropy and random forest) were applied to a heterogeneous section of seabed on Rockall Bank, NE Atlantic, for which landscape indices describing the spatial arrangement of habitat patches were calculated. The predictive maps were based on remotely operated vehicle (ROV) imagery transects high-resolution autonomous underwater vehicle (AUV) sidescan backscatter maps. Area under the curve (AUC) and accuracy indicated similar performances for the three models tested, but performance varied by species assemblage, with the transitional species assemblage showing the weakest predictive performances. Spatial predictions of habitat suitability differed between statistical approaches, but niche similarity metrics showed redundancy analysis and random forest predictions to be most similar. As one statistical technique could not be found to outperform the others when all assemblages were considered, ensemble mapping techniques, where the outputs of many models are combined, were applied. They showed higher accuracy than any single model. Different statistical approaches for predictive habitat modelling possess varied strengths and weaknesses and by examining the outputs of a range of modelling techniques and their differences, more robust predictions, with better described variation and areas of uncertainties, can be achieved. As improvements to prediction outputs can be achieved without additional costly data collection, ensemble mapping approaches have clear value for spatial management.
Article
Full-text available
Understanding and predicting the responses of wide-ranging marine predators such as cetaceans, seabirds, sharks, turtles, pinnipeds and large migratory fish to dynamic oceanographic conditions requires habitat-based models that can sufficiently capture their environmental preferences. Marine ecosystems are inherently dynamic, and animal–environment interactions are known to occur over multiple, nested spatial and temporal scales. The spatial resolution and temporal averaging of environmental data layers are therefore key considerations in modelling the environmental determinants of habitat selection. The utility of environmental data contemporaneous to animal presence or movement (e.g. daily, weekly), versus synoptic products (monthly, seasonal, climatological) is currently debated, as are the trade-offs between near real-time, high resolution and composite (i.e. synoptic, cloud-free) data fields.
Article
Full-text available
sdm is an object-oriented, reproducible and extensible, platform for species distribution modelling. It uses individual species and community-based approaches, enabling ensembles of models to be fitted and evaluated, to project species potential distributions in space and time. It provides a standardized and unified structure for handling species distributions data and modelling techniques, and supports markedly different modelling approaches, including correlative, process-based (mechanistic), agent-based, and cellular automata. The object-oriented design of software is such that scientists can modify existing methods, extend the framework by developing new methods or modelling procedures, and share them to be reproduced by other scientists. sdm can handle spatial and temporal data for single or multiple species and uses high performance computing solutions to speed up modelling and simulations. The framework is implemented in R, providing a flexible and easy-to-use GUI interface.
Article
Full-text available
How complex does a model need to be to provide useful predictions is a matter of continuous debate across environmental sciences. In the species distributions modelling literature, studies have demonstrated that more complex models tend to provide better fits. However, studies have also shown that predictive performance does not always increase with complexity. Testing of species distributions models is challenging because independent data for testing are often lacking, but a more general problem is that model complexity has never been formally described in such studies. Here, we systematically examine predictive performance of models against data and models of varying complexity. We introduce the concept of computational complexity, widely used in theoretical computer sciences, to quantify model complexity. In addition, complexity of species distributional data is characterized by their geometrical properties. Tests involved analysis of models’ ability to predict virtual species distributions in the same region and the same time as used for training the models, and to project distributions in different times under climate change. Of the eight species distribution models analyzed five (Random Forest, boosted regression trees, generalized additive models, multivariate adaptive regression splines, MaxEnt) showed similar performance despite differences in computational complexity. The ability of models to forecast distributions under climate change was also not affected by model complexity. In contrast, geometrical characteristics of the data were related to model performance in several ways: complex datasets were consistently more difficult to model, and the complexity of the data was affected by the choice of predictors and the type of data analyzed. Given our definition of complexity, our study contradicts the widely held view that the complexity of species distributions models has significant effects in their predictive ability while findings support for previous observations that the properties of species distributions data and their relationship with the environment are strong predictors of model success.
Article
Full-text available
Spatially explicit information on species distributions for conservation planning is invariably incomplete, and therefore requires the use of surrogates to represent broad-scale patterns of biodiversity. Despite significant interest in the effectiveness of surrogates for predicting biodiversity, few studies explore questions involving the ability of surrogates to ‘incidentally’ represent ‘unknown’ features of conservation interest. Here, we use the Great Barrier Reef marine reserve network to examine factors affecting incidental representation of conservation features that were unknown at the time the reserve network was established. We examine how incidental representation of 39 seabed habitats and biological assemblages was affected by the spatial characteristics of the features, the conservation objectives, the spatial configuration of no-take areas, and the opportunity cost of conservation. Our results show cost was closely correlated to incidental representation. However, incidental representation was achieved, even in a region with only coarse-scale environmental data, by adopting a precautionary approach that explicitly considered the potential for unknown features. Our results indicate that incidental representation is enhanced by partitioning selection units along biophysical gradients to account for unknown within-feature variability and ensuring that no-take areas are well-distributed throughout the region; setting high conservation objectives in order to maximize the chances of capturing unknown features incidentally; and carefully considering the designation of cost to planning units when using decision-support tools for reserve design. The lessons learned from incidental representation in the Great Barrier Reef have implications for conservation planning in other regions, particularly those that lack detailed environmental and ecological data.This article is protected by copyright. All rights reserved
Article
Full-text available
Species data held in museum and herbaria, survey data and opportunistically observed data are a substantial information resource. A key challenge in using these data is the uncertainty about where an observation is located. This is important when the data are used for species distribution modelling (SDM), because the coordinates are used to extract the environmental variables and thus, positional error may lead to inaccurate estimation of the species–environment relationship. The magnitude of this effect is related to the level of spatial autocorrelation in the environmental variables. Using local spatial association can be relevant because it can lead to the identification of the specific occurrence records that cause the largest drop in SDM accuracy. Therefore, in this study, we tested whether the SDM predictions are more affected by positional uncertainty originating from locations that have lower local spatial association in their predictors. We performed this experiment for Spain and the Netherlands, using simulated datasets derived from well known species distribution models (SDMs). We used the K statistic to quantify the local spatial association in the predictors at each species occurrence location. A probabilistic approach using Monte Carlo simulations was employed to introduce the error in the species locations. The results revealed that positional uncertainty in species occurrence data at locations with low local spatial association in predictors reduced the prediction accuracy of the SDMs. We propose that local spatial association is a way to identify the species occurrence records that require treatment for positional uncertainty. We also developed and present a tool in the R environment to target observations that are likely to create error in the output from SDMs as a result of positional uncertainty.
Article
Full-text available
Coral reef ecosystems, the most varied on earth, continually face destruction from anthropogenic and natural threats. The U.S. Coral Reef Task Force seeks to characterize and map priority coral reef ecosystems in the U.S./Trust Territories by 2009. Building upon NOAA Biogeography shallow-water classifications based on Ikonos imagery, presented here are new methods, based on acoustic data, for classifying benthic terrain below 30 m, around Tutuila, American Samoa. The result is a new classification scheme for American Samoa that extends and improves the NOAA Biogeography scheme, which, although developed for Pacific island nations and territories, is only applicable to a maximum depth of 30 m, due to the limitations of satellite imagery. The scheme may be suitable for developing habitat maps pinpointing high biodiversity around coral reefs throughout the western Pacific.
Article
Full-text available
Multibeam surveys can provide detailed bathymetry data for the continental slope from which quantitative descriptors of the seabed terrain (e.g., slope) may be obtained. We illustrate the value of these descriptors for benthic habitat mapping, and highlight the advantages of multiscale analysis. We examine the application of these descriptors as predictor variables for species distribution models, which are particularly valuable in the deep sea where opportunities to directly survey the benthic fauna remain limited. Our initial models are encouraging and suggest that wider adoption of these methods may assist the delivery of ecologically relevant information to marine resource managers.
Article
Full-text available
Aim To examine the impacts of climate change on endemic birds, which are of global significance for conservation, on a continent with few such assessments. We specifically assess projected range changes in relation to the Important Bird Areas (IBAs) network and assess the possible consequences for conservation. Location South Africa, Lesotho and Swaziland. Methods The newly emerging ensemble modelling approach is used with 50 species, four climate change models for the period 2070–2100 and eight bioclimatic niche models in the statistical package biomod. Model evaluation is done using the receiver operating characteristic and the recently introduced true skill statistic. Future projections are made considering two extreme assumptions: species have full dispersal ability and species have no dispersal ability. A consensus forecast is identified using principal components analysis. This forecast is interpreted in terms of the IBA network. An irreplaceability analysis is used to highlight priority IBAs for conservation attention in terms of climate change. Results The majority of species (62%) are predicted to lose climatically suitable space. Five species lose at least 85% of their climatically suitable space. Many IBAs lose species (41%; 47 IBAs) and show high rates of species turnover of more than 50% (77%; 95 IBAs). Highly irreplaceable regions for endemic species become highly localized under climate change, meaning that the endemic species analysed here experience similar range contractions to maintain climate niches. Main conclusions The South African IBAs network is likely to become less effective for conserving endemic birds under climate change. The irreplaceability analysis identified key refugia for endemic species under climate change, but many of these areas are not currently IBAs. In addition, many of these high-priority areas that are IBAs fall outside the current formal protected areas network.
Article
Full-text available
Prediction of species’ distributions is central to diverse applications in ecology, evolution and conservation science. There is increasing electronic access to vast sets of occurrence records in museums and herbaria, yet little effective guidance on how best to use this information in the context of numerous approaches for modelling distributions. To meet this need, we compared 16 modelling methods over 226 species from 6 regions of the world, creating the most comprehensive set of model comparisons to date. We used presence-only data to fit models, and independent presence-absence data to evaluate the predictions. Along with well-established modelling methods such as generalised additive models and GARP and BIOCLIM, we explored methods that either have been developed recently or have rarely been applied to modelling species’ distributions. These include machine-learning methods and community models, both of which have features that may make them particularly well suited to noisy or sparse information, as is typical of species’ occurrence data. Presence-only data were effective for modelling species’ distributions for many species and regions. The novel methods consistently outperformed more established methods. The results of our analysis are promising for the use of data from museums and herbaria, especially as methods suited to the noise inherent in such data improve.
Article
Full-text available
In the last two decades, interest in species distribution models (SDMs) of plants and animals has grown dramatically. Recent advances in SDMs allow us to potentially forecast anthropogenic effects on patterns of biodiversity at different spatial scales. However, some limitations still preclude the use of SDMs in many theoretical and practical applications. Here, we provide an overview of recent advances in this field, discuss the ecological principles and assumptions underpinning SDMs, and highlight critical limitations and decisions inherent in the construction and evaluation of SDMs. Particular emphasis is given to the use of SDMs for the assessment of climate change impacts and conservation management issues. We suggest new avenues for incorporating species migration, population dynamics, biotic interactions and community ecology into SDMs at multiple spatial scales. Addressing all these issues requires a better integration of SDMs with ecological theory.
Article
Full-text available
The western rock lobster, Panulirus cygnus, is endemic to Western Australia and supports substantial commercial and recreational fisheries. Due to and its wide distribution and the commercial and recreational importance of the species a key component of managing western rock lobster is understanding the ecological processes and interactions that may influence lobster abundance and distribution. Using terrain analyses and distribution models of substrate and benthic biota, we assess the physical drivers that influence the distribution of lobsters at a key fishery site. Using data collected from hydroacoustic and towed video surveys, 20 variables (including geophysical, substrate and biota variables) were developed to predict the distributions of substrate type (three classes of reef, rhodoliths and sand) and dominant biota (kelp, sessile invertebrates and macroalgae) within a 40 km(2) area about 30 km off the west Australian coast. Lobster presence/absence data were collected within this area using georeferenced pots. These datasets were used to develop a classification tree model for predicting the distribution of the western rock lobster. Interestingly, kelp and reef were not selected as predictors. Instead, the model selected geophysical and geomorphic scalar variables, which emphasise a mix of terrain within limited distances. The model of lobster presence had an adjusted D(2) of 64 and an 80% correct classification. Species distribution models indicate that juxtaposition in fine scale terrain is most important to the western rock lobster. While key features like kelp and reef may be important to lobster distribution at a broad scale, it is the fine scale features in terrain that are likely to define its ecological niche. Determining the most appropriate landscape configuration and scale will be essential to refining niche habitats and will aid in selecting appropriate sites for protecting critical lobster habitats.
Article
Full-text available
Aim Various statistical techniques have been used to model species probabilities of occurrence in response to environmental conditions. This paper provides a comprehensive assessment of methods and investigates whether errors in model predictions are associated to specific kinds of geographical and environmental distributions of species. Location Portugal, Western Europe. Methods Probabilities of occurrence for 44 species of amphibians and reptiles in Portugal were modelled using seven modelling techniques: Gower metric, Ecological Niche Factor Analysis, classification trees, neural networks, generalized linear models, generalized additive models and spatial interpolators. Generalized linear and additive models were constructed with and without a term accounting for spatial autocorrelation. Model performance was measured using two methods: sensitivity and Kappa index. Species were grouped according to their spatial (area of occupancy and extent of occurrence) and environmental (marginality and tolerance) distributions. Two-way comparison tests were performed to detect significant interactions between models and species groups. Results Interaction between model and species groups was significant for both sensitivity and Kappa index. This indicates that model performance varied for species with different geographical and environmental distributions. Artificial neural networks performed generally better, immediately followed by generalized additive models including a covariate term for spatial autocorrelation. Non-parametric methods were preferred to parametric approaches, especially when modelling distributions of species with a greater area of occupancy, a larger extent of occurrence, lower marginality and higher tolerance. Main conclusions This is a first attempt to relate performance of modelling techniques with species spatial and environmental distributions. Results indicate a strong relationship between model performance and the kinds of species distributions being modelled. Some methods performed generally better, but no method was superior in all circumstances. A suggestion is made that choice of the appropriate method should be contingent on the goals and kinds of distributions being modelled.
Article
Full-text available
Aim Species distribution models (SDMs) have been used to address a wide range of theoretical and applied questions in the terrestrial realm, but marine-based applications remain relatively scarce. In this review, we consider how conceptual and practical issues associated with terrestrial SDMs apply to a range of marine organisms and highlight the challenges relevant to improving marine SDMs. Location We include studies from both marine and terrestrial systems that encompass many geographic locations around the globe. Methods We first performed a literature search and analysis of marine and terrestrial SDMs in ISI Web of Science to assess trends and applications. Using knowledge from terrestrial applications, we critically evaluate the application of SDMs in marine systems in the context of ecological factors (dispersal, species interactions, aggregation and ontogenetic shifts) and practical considerations (data quality, alternative modelling approaches and model validation) that facilitate or create difficulties for model application. Results The relative importance of ecological factors to be considered when applying SDMs varies among terrestrial and marine organisms. Correctly incorporating dispersal is frequently considered an important issue for terrestrial models, but because there is greater potential for dispersal in the ocean, it is often less of a concern in marine SDMs. By contrast, ontogenetic shifts and feeding have received little attention in terrestrial SDM applications, but these factors are important to many marine SDMs. Opportunities also exist for applying more advanced SDM approaches in the marine realm, including mechanistic ecophysiological models, where water balance and heat transfer equations are simpler for some marine organisms relative to their terrestrial counterparts. Main conclusions SDMs have generally been under-utilized in the marine realm relative to terrestrial applications. Correlative SDM methods should be tested on a range of marine organisms, and we suggest further development of methods that address ontogenetic shifts and feeding interactions. We anticipate developments in, and cross-fertilization between, coupled correlative and process-based SDMs, mechanistic eco-physiological SDMs, and spatial population dynamic models for climate change and species invasion applications in particular. Comparisons of the outputs of different model types will provide insight that is useful for improved spatial management of marine species.
Article
Research on spatial data analysis has developed a number of local indicators of spatial association (LISA), which allow exploration of local patterns in spatial data. These include local Moran's [Formula presented] and local Geary's [Formula presented], as well as [Formula presented] and [Formula presented] that can be used for continuous or interval variables only. Despite numerous situations where qualitative (nominal/categorical) variables are encountered, few attempts have been devoted to the development of methods to explore the local spatial pattern in categorical data. To our knowledge, there is no indicator of local spatial association that can be used for both continuous and categorical data at the same time. In this paper, we propose a new local indicator of spatial association, called the entropy-based local indicator of spatial association (ELSA), that can be used for both categorical and continuous spatial data. ELSA quantifies the degree of spatial association of a variable at each location relative to the same variable at the neighbouring locations. This indicator simultaneously incorporates both spatial and attribute aspects of spatial association into account. The values of ELSA vary between 0 and 1, which denote highest and lowest spatial association, respectively. We compare ELSA to existing statistics such as Local Moran's I and test the power and size of the new statistic. We also introduce the ”entrogram” a novel approach for exploring the global spatial structure within the entire area (like a variogram). This study showed that the ELSA is consistent and robust, and is therefore suitable for applications in a wide range of disciplines. The ELSA algorithm is made available as an R-package (elsa).
Book
The methodology used to construct tree structured rules is the focus of this monograph. Unlike many other statistical procedures, which moved from pencil and paper to calculators, this text's use of trees was unthinkable before computers. Both the practical and theoretical sides have been developed in the authors' study of tree methods. Classification and Regression Trees reflects these two sides, covering the use of trees as a data analysis method, and in a more mathematical framework, proving some of their fundamental properties.
Article
Kroodsma et al. (Reports, 23 February 2018, p. 904) mapped the global footprint of fisheries. Their estimates of footprint and resulting contrasts between the scale of fishing and agriculture are an artifact of the spatial scale of analysis. Reanalyses of their global (all vessels) and regional (trawling) data at higher resolution reduced footprint estimates by factors of >10 and >5, respectively.
Article
Numerous machine-learning classifiers are available for benthic habitat map production, which can lead to different results. This study highlights the performance of the Random Forest (RF) classifier, which was significantly better than Classification Trees (CT), Naïve Bayes (NB), and a multi-model ensemble in terms of overall accuracy, Balanced Error Rate (BER), Kappa, and area under the curve (AUC) values. RF accuracy was often higher than 90% for each substratum class, even at the most detailed level of the substratum classification and AUC values also indicated excellent performance (0.8–1). Total agreement between classifiers was high at the broadest level of classification (75–80%) when differentiating between hard and soft substratum. However, this sharply declined as the number of substratum categories increased (19–45%) including a mix of rock, gravel, pebbles, and sand. The model ensemble, produced from the results of all three classifiers by majority voting, did not show any increase in predictive performance when compared to the single RF classifier. This study shows how a single classifier may be sufficient to produce benthic seabed maps and model ensembles of multiple classifiers.
Article
To evaluate the conservation performance of Australia's continental-scale network of marine reserves for deep-water octocorals using three criteria: (1) Representation: what fraction of the sampled deep-water octocoral fauna within Australia's marine jurisdiction is contained in reserves?; (2) Species turnover: to what degree do reserves share species with non-protected areas?; and (3) Biodiversity: do existing reserves spatially maximize protection of species richness? Australia's continental exclusive economic zone and Norfolk Island region. A new Australia-wide, taxonomically consistent, dataset for distributions of deep-water octocorals (> 80 m depth) is used to calculate metrics of reserve performance for Australia's national marine reserve network. Australia's reserve network represents 52% (270 of 518 species) of deep-water octocoral species. Coral representation is roughly proportional to spatial extent of reserves in different regions. The proportion of species with joint distributions in reserves and non-reserves ranges between 0% and 27% of the regional species pool; this may be a reflection of the high species turnover between sampling sites, and to some extent may also reflect sparse sampling coverage. Sites inside reserves have comparable species richness, contain similar numbers of rare species and do not differ in taxonomic diversity to sites outside reserves. Australia's marine reserve network is a significant investment in conservation. The first broad biological evaluation of the network's performance, using deep-water octocorals, indicates that whilst the level of species representation is broadly proportional to the network's spatial coverage, reserves do not appear to be areas of significantly higher octocoral biodiversity compared to areas outside reserves and that there is low species representation within highly protected (IUCN I or II) zones. Considering constraints on reserve expansion and spatial re-configuration, ‘off-reserve’ management tools (e.g. fishery closures, environmentally responsible practices by extractive industries) are likely to be important to enhance overall conservation outcomes.
Article
Species distribution models (SDMs) are widely used to explain and predict species ranges and environmental niches. They are most commonly constructed by inferring species' occurrence–environment relationships using statistical and machine-learning methods. The variety of methods that can be used to construct SDMs (e.g. generalized linear/additive models, tree-based models, maximum entropy, etc.), and the variety of ways that such models can be implemented, permits substantial flexibility in SDM complexity. Building models with an appropriate amount of complexity for the study objectives is critical for robust inference. We characterize complexity as the shape of the inferred occurrence–environment relationships and the number of parameters used to describe them, and search for insights into whether additional complexity is informative or superfluous. By building ‘under fit’ models, having insufficient flexibility to describe observed occurrence–environment relationships, we risk misunderstanding the factors shaping species distributions. By building ‘over fit’ models, with excessive flexibility, we risk inadvertently ascribing pattern to noise or building opaque models. However, model selection can be challenging, especially when comparing models constructed under different modeling approaches. Here we argue for a more pragmatic approach: researchers should constrain the complexity of their models based on study objective, attributes of the data, and an understanding of how these interact with the underlying biological processes. We discuss guidelines for balancing under fitting with over fitting and consequently how complexity affects decisions made during model building. Although some generalities are possible, our discussion reflects differences in opinions that favor simpler versus more complex models. We conclude that combining insights from both simple and complex SDM building approaches best advances our knowledge of current and future species ranges.
Conference Paper
Australias Integrated Marine Observing System (IMOS) has a strategic focus on the impact of major boundary currents on continental shelf environments, ecosystems and biodiversity. To improve our understanding of natural, climate change, and human-induced variability in shelf environments, the IMOS Autonomous Underwater Vehicle (AUV) facility has been charged with generating physical and biological observations of benthic variables that cannot be cost-effectively obtained by other means. Starting in 2010, the IMOS AUV facility began collecting precisely navigated benthic imagery using AUVs at selected reference sites on Australias shelf. This observing program capitalizes on the unique capabilities of AUVs that have allowed repeated visits to the reference sites, providing a critical observational link between oceanographic and benthic processes. This paper provides a brief overview of the relevant capabilities of the AUV facility, the design of the IMOS benthic sampling program, and some preliminary results. We also report on some of the challenges and potential benefits to be realized from a benthic observation system that collects several TB of geo-referenced stereo imagery a year. This includes collaborative semi-automated image analysis, clustering and classification, large scale visualization and data mining, and lighting correction for change detection and characterization. We also mention some of the lessons from operating an AUV-based monitoring program and future work in this area.
Article
1. As systems of marine protected areas (MPAs) expand globally, there is a risk that new MPAs will be biased toward places that are remote or unpromising for extractive activities, and hence follow the trend of terrestrial protected areas in being ‘residual’ to commercial uses. Such locations typically provide little protection to the species and ecosystems that are most exposed to threatening processes. 2. There are strong political motivations to establish residual reserves that minimize costs and conflicts with users of natural resources. These motivations will likely remain in place as long as success continues to be measured in terms of area (km2) protected. 3. The global pattern of MPAs was reviewed and appears to be residual, supported by a rapid growth of large, remote MPAs. The extent to which MPAs in Australia are residual nationally and also regionally within the Great Barrier Reef (GBR) Marine Park was also examined. 4. Nationally, the recently announced Australian Commonwealth marine reserves were found to be strongly residual, making almost no difference to ‘business as usual’ for most ocean uses. Underlying this result was the imperative to minimize costs, but without the spatial constraints of explicit quantitative objectives for representing bioregions or the range of ecological features in highly protected zones. 5. In contrast, the 2004 rezoning of the GBR was exemplary, and the potential for residual protection was limited by applying a systematic set of planning principles, such as representing a minimum percentage of finely subdivided bioregions. Nonetheless, even at this scale, protection was uneven between bioregions. Within-bioregion heterogeneity might have led to no-take zones being established in areas unsuitable for trawling with a risk that species assemblages differ between areas protected and areas left available for trawling. 6. A simple four-step framework of questions for planners and policy makers is proposed to help reverse the emerging residual tendency of MPAs and maximize their effectiveness for conservation. This involves checks on the least-cost approach to establishing MPAs in order to avoid perverse outcomes.
Article
With the rise of new powerful statistical techniques and GIS tools, the development of predictive habitat distribution models has rapidly increased in ecology. Such models are static and probabilistic in nature, since they statistically relate the geographical distribution of species or communities to their present environment. A wide array of models has been developed to cover aspects as diverse as biogeography, conservation biology, climate change research, and habitat or species management. In this paper, we present a review of predictive habitat distribution modeling. The variety of statistical techniques used is growing. Ordinary multiple regression and its generalized form (GLM) are very popular and are often used for modeling species distributions. Other methods include neural networks, ordination and classification methods, Bayesian models, locally weighted approaches (e.g. GAM), environmental envelopes or even combinations of these models. The selection of an appropriate method should not depend solely on statistical considerations. Some models are better suited to reflect theoretical findings on the shape and nature of the species’ response (or realized niche). Conceptual considerations include e.g. the trade-off between optimizing accuracy versus optimizing generality. In the field of static distribution modeling, the latter is mostly related to selecting appropriate predictor variables and to designing an appropriate procedure for model selection. New methods, including threshold-independent measures (e.g. receiver operating characteristic (ROC)-plots) and resampling techniques (e.g. bootstrap, cross-validation) have been introduced in ecology for testing the accuracy of predictive models. The choice of an evaluation measure should be driven primarily by the goals of the study. This may possibly lead to the attribution of different weights to the various types of prediction errors (e.g. omission, commission or confusion). Testing the model in a wider range of situations (in space and time) will permit one to define the range of applications for which the model predictions are suitable. In turn, the qualification of the model depends primarily on the goals of the study that define the qualification criteria and on the usability of the model, rather than on statistics alone.
Article
Species distributions models (SDMs) are commonly used to assess potential species' range shifts or extinction risk under climate change. It has been suggested that the use of ensemble forecasts, where a variety of model algorithms are used to generate consensus predictions, are preferred to individual SDMs by avoiding bias or prediction error inherent in a single modeling approach. Whereas several studies have assessed the performance of ensemble predictions using cross-validation or data-partitioning approaches, few studies have assessed the predictive accuracy of ensemble forecasts under climate change by using temporally independent model validation data. We used five SDM approaches to develop consensus forecasts of distributions of 145 vascular plant species from California in the 1930s and tested their projections against current distributions, a span of approximately 75 years. When evaluated with a portion of the model training data, consensus forecasts were highly accurate with an average AUC value of 0.97. False positive and false negative error rates were also low, exhibiting similar performance to random forest models. However, when evaluated with temporally independent data, the accuracy of consensus forecasts was similar to that of generalized linear and generalized additive models, with an average AUC value of 0.83. Our results suggest that the high levels of predictive accuracy exhibited by consensus forecasts when using data partitioning approaches may not reflect their performance when predicting temporally independent data. We contend that consensus forecasts may not represent the best approach for predicting species distributions under future climatic change, as they may not provide superior predictive accuracy in novel temporal domains compared to traditional modeling approaches that more readily lend themselves to ecological interpretation of model structure.
Article
Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest. The generalization error for forests converges a.s. to a limit as the number of trees in the forest becomes large. The generalization error of a forest of tree classifiers depends on the strength of the individual trees in the forest and the correlation between them. Using a random selection of features to split each node yields error rates that compare favorably to Adaboost (Y. Freund & R. Schapire, Machine Learning: Proceedings of the Thirteenth International conference, ***, 148–156), but are more robust with respect to noise. Internal estimates monitor error, strength, and correlation and these are used to show the response to increasing the number of features used in the splitting. Internal estimates are also used to measure variable importance. These ideas are also applicable to regression.
Article
Habitat models for animal species are important tools in conservation planning. We assessed the need to consider several scales in a case study for three amphibian and two grasshopper species in the post-mining landscapes near Leipzig (Germany). The two species groups were selected because habitat analyses for grasshoppers are usually conducted on one scale only whereas amphibians are thought to depend on more than one spatial scale.First, we analysed how the preference to single habitat variables changed across nested scales. Most environmental variables were only significant for a habitat model on one or two scales, with the smallest scale being particularly important. On larger scales, other variables became significant, which cannot be recognized on lower scales. Similar preferences across scales occurred in only 13 out of 79 cases and in 3 out of 79 cases the preference and avoidance for the same variable were even reversed among scales.Second, we developed habitat models by using a logistic regression on every scale and for all combinations of scales and analysed how the quality of habitat models changed with the scales considered. To achieve a sufficient accuracy of the habitat models with a minimum number of variables, at least two scales were required for all species except for Bufo viridis, for which a single scale, the microscale, was sufficient. Only for the European tree frog (Hyla arborea), at least three scales were required.The results indicate that the quality of habitat models increases with the number of surveyed variables and with the number of scales, but costs increase too. Searching for simplifications in multi-scaled habitat models, we suggest that 2 or 3 scales should be a suitable trade-off, when attempting to define a suitable microscale.
Article
For many applications the continuous prediction afforded by species distribution modeling must be converted to a map of presence or absence, so a threshold probability indicative of species presence must be fixed. Because of the bias in probability outputs due to frequency of presences (prevalence), a fixed threshold value, such as 0.5, does not usually correspond to the threshold above which the species is more likely to be present. In this paper four threshold criteria are compared for a wide range of sample sizes and prevalences, modeling a virtual species in order to avoid the omnipresent error sources that the use of real species data implies. In general, sensitivity–specificity difference minimizer and sensitivity–specificity sum maximizer criteria produced the most accurate predictions. The widely-used 0.5 fixed threshold and Kappa-maximizer criteria are the worst ones in almost all situations. Nevertheless, whatever the criteria used, the threshold value chosen and the research goals that determined its choice must be stated.
Article
Simple interval estimate methods for proportions exhibit poor coverage and can produce evidently inappropriate intervals. Criteria appropriate to the evaluation of various proposed methods include: closeness of the achieved coverage probability to its nominal value; whether intervals are located too close to or too distant from the middle of the scale; expected interval width; avoidance of aberrations such as limits outside [0,1] or zero width intervals; and ease of use, whether by tables, software or formulae. Seven methods for the single proportion are evaluated on 96,000 parameter space points. Intervals based on tail areas and the simpler score methods are recommended for use. In each case, methods are available that aim to align either the minimum or the mean coverage with the nominal 1 - α.
Article
Aim Spatial modelling techniques are increasingly used in species distribution modelling. However, the implemented techniques differ in their modelling performance, and some consensus methods are needed to reduce the uncertainty of predictions. In this study, we tested the predictive accuracies of five consensus methods, namely Weighted Average (WA), Mean(All), Median(All), Median(PCA), and Best, for 28 threatened plant species. Location North-eastern Finland, Europe. Methods The spatial distributions of the plant species were forecasted using eight state-of-the-art single-modelling techniques providing an ensemble of predictions. The probability values of occurrence were then combined using five consensus algorithms. The predictive accuracies of the single-model and consensus methods were assessed by computing the area under the curve (AUC) of the receiver-operating characteristic plot. Results The mean AUC values varied between 0.697 (classification tree analysis) and 0.813 (random forest) for the single-models, and from 0.757 to 0.850 for the consensus methods. WA and Mean(All) consensus methods provided significantly more robust predictions than all the single-models and the other consensus methods. Main conclusions Consensus methods based on average function algorithms may increase significantly the accuracy of species distribution forecasts, and thus they show considerable promise for different conservation biological and biogeographical applications.
Article
High resolution remote sensing data facilitate the use of small-scale habitat features such as trees or hedges in the analysis of species–habitat relationships. Such data potentially enable more accurate species–habitat mapping than lower resolution data. Here, for the first time, we systematically investigated this hypothesis by altering the spatial resolution from 1 m up to 1000 m grain size in species–habitat models of 13 bird species. The study area covered the Nidda river catchment in central Germany, a large heterogeneous landscape of 1620 km2. A high resolution habitat map of the area was converted to coarser spatial and thematic resolutions in seven steps. We investigated how model performance responded to grain size, and we compared the differential effects of spatial resolution and thematic resolution on model performance. Explained deviance (D2) of the bird models generally decreased with coarser spatial resolution of the data, although it did not decrease monotonically in all species. On average across all species, model D2 decreased from 41.5 at 1 m grain size to 15.9 at 1000 m grain size. Ten species were best modelled at 1 m, two species at 3 m and one species at 32 m grain size. Model performance degraded continuously with increasing grain size, both in habitat generalist and habitat specialist bird species, and was systematically lower in habitat generalists. The higher model performance observed at finer grain sizes was most likely caused by the combination of three factors: (1) high spatial accuracy of bird records and (2) a more precise location and delineation of habitat features and, (3) to a lesser degree, by more habitat types differentiated in maps of finer resolution. We conclude that higher spatial and thematic resolution data can be essential for deriving accurate predictions on bird distribution patterns from species–habitat models. Especially for bird species that are sensitive to specific land-use types or to small-scaled habitat features, a grain size of 1–3 m seems most promising.
Article
Predictive species distribution modelling (SDM) has become an essential tool in biodiversity conservation and management. The choice of grain size (resolution) of environmental layers used in modelling is one important factor that may affect predictions. We applied 10 distinct modelling techniques to presence-only data for 50 species in five different regions, to test whether: (1) a 10-fold coarsening of resolution affects predictive performance of SDMs, and (2) any observed effects are dependent on the type of region, modelling technique, or species considered. Results show that a 10 times change in grain size does not severely affect predictions from species distribution models. The overall trend is towards degradation of model performance, but improvement can also be observed. Changing grain size does not equally affect models across regions, techniques, and species types. The strongest effect is on regions and species types, with tree species in the data sets (regions) with highest locational accuracy being most affected. Changing grain size had little influence on the ranking of techniques: boosted regression trees remain best at both resolutions. The number of occurrences used for model training had an important effect, with larger sample sizes resulting in better models, which tended to be more sensitive to grain. Effect of grain change was only noticeable for models reaching sufficient performance and/ or with initial data that have an intrinsic error smaller than the coarser grain size.
Article
Knowledge of landscape spatial patterns of seagrasses and their rates of loss and natural colonization is critical for understanding the ecology of this group of submerged aquatic plants. Seagrasses form extensive meadows that occupy sheltered coastal seas of the world. In this paper, we examine the multi-scale variability of three seagrass species over a large near-shore region (42 km2) in Western Australia. Geostatistical non-parametric methods were used to explore spatial variation in presence of Amphibolis griffithii, Posidonia coriacea and P. sinuosa, and to identify the spatial scales at which distinct patterns in the species distributions occur: <50, 50–610 and >610 m. Each species showed unique variance structure across local (<50 and 50–610 m) and regional scales (>610 m), suggesting differences in species biology, environmental requirements, inter-species interactions, and their ability to modify their environment. These observations reflect that 1) seagrass landscapes are created by processes that independently act on each seagrass species at different spatial scales; 2) the species’ distributions differ in their hydrodynamic forcing, and; 3) seagrass species distributions reflect colonization history such that related species are separated in space because they have different places in the successional sequence. This cross-scale study demonstrates that shoot studies only partly address the spatial structure of seagrass landscapes and further large-scale spatially-explicit research is required before we can interpret the driving processes.
Article
Summary 1. In recent years the use of species distribution models by ecologists and conservation managers has increased considerably, along with an awareness of the need to provide accuracy assessment for predictions of such models. The kappa statistic is the most widely used measure for the performance of models generating presence-absence predictions, but several studies have criticized it for being inherently dependent on prevalence, and argued that this dependency introduces statistical artefacts to estimates of predictive accuracy. This criticism has been supported recently by computer simulations showing that kappa responds to the prevalence of the modelled species in a unimodal fashion. 2. In this paper we provide a theoretical explanation for the observed dependence of kappa on prevalence, and introduce into ecology an alternative measure of accuracy, the true skill statistic (TSS), which corrects for this dependence while still keeping all the advantages of kappa. We also compare the responses of kappa and TSS to prevalence using empirical data, by modelling distribution patterns of 128 species of woody plant in Israel. 3. The theoretical analysis shows that kappa responds in a unimodal fashion to vari- ation in prevalence and that the level of prevalence that maximizes kappa depends on the ratio between sensitivity (the proportion of correctly predicted presences) and specificity (the proportion of correctly predicted absences). In contrast, TSS is inde- pendent of prevalence. 4. When the two measures of accuracy were compared using empirical data, kappa showed a unimodal response to prevalence, in agreement with the theoretical analysis. TSS showed a decreasing linear response to prevalence, a result we interpret as reflecting true ecological phenomena rather than a statistical artefact. This interpretation is supported by the fact that a similar pattern was found for the area under the ROC curve, a measure known to be independent of prevalence. 5. Synthesis and applications . Our results provide theoretical and empirical evidence that kappa, one of the most widely used measures of model performance in ecology, has serious limitations that make it unsuitable for such applications. The alternative we suggest, TSS, compensates for the shortcomings of kappa while keeping all of its advantages. We therefore recommend the TSS as a simple and intuitive measure for the performance of species distribution models when predictions are expressed as presence- absence maps.
Article
Despite two centuries of exploration, our understanding of factors determining the distribution of life on Earth is in many ways still in its infancy. Much of the disagreement about governing processes of variation in species richness may be the result of differences in our perception of species-richness patterns. Until recently, most studies of large-scale species-richness patterns assumed implicitly that patterns and mechanisms were scale invariant. Illustrated with examples and a quantitative analysis of published data on altitudinal gradients of species richness (n = 204), this review discusses how scale effects (extent and grain size) can influence our perception of patterns and processes. For example, a hump-shaped altitudinal species-richness pattern is the most typical (c. 50%), with a monotonic decreasing pattern (c. 25%) also frequently reported, but the relative distribution of patterns changes readily with spatial grain and extent. If we are to attribute relative impact to various factors influencing species richness and distribution and to decide at which point along a spatial and temporal continuum they act, we should not ask only how results vary as a function of scale but also search for consistent patterns in these scale effects. The review concludes with suggestions of potential routes for future analytical exploration of species-richness patterns.
Article
Species distribution models (SDMs) are numerical tools that combine observations of species occurrence or abundance with environmental estimates. They are used to gain ecological and evolutionary insights and to predict distributions across landscapes, sometimes requiring extrapolation in space and time. SDMs are now widely used across terrestrial, freshwater, and marine realms. Differences in methods between disciplines reflect both differences in species mobility and in “established use. Model realism and robustness is influenced by selection of relevant predictors and modeling method, consideration of scale, how the interplay between environmental and geographic factors is handled, and the extent of extrapolation. Current linkages between SDM practice and ecological theory are often weak, hindering progress. Remaining challenges include: improvement of methods for modeling presence-only data and for model selection and evaluation; accounting for biotic interactions; and assessing model uncertainty.
Article
The availability of detailed environmental data, together with inexpensive and powerful computers, has fueled a rapid increase in predictive modeling of species environmental requirements and geographic distributions. For some species, detailed presence/absence occurrence data are available, allowing the use of a variety of standard statistical techniques. However, absence data are not available for most species. In this paper, we introduce the use of the maximum entropy method (Maxent) for modeling species geographic distributions with presence-only data. Maxent is a general-purpose machine learning method with a simple and precise mathematical formulation, and it has a number of aspects that make it well-suited for species distribution modeling. In order to investigate the efficacy of the method, here we perform a continental-scale case study using two Neotropical mammals: a lowland species of sloth, Bradypus variegatus, and a small montane murid rodent, Microryzomys minutus. We compared Maxent predictions with those of a commonly used presence-only modeling method, the Genetic Algorithm for Rule-Set Prediction (GARP). We made predictions on 10 random subsets of the occurrence records for both species, and then used the remaining localities for testing. Both algorithms provided reasonable estimates of the species’ range, far superior to the shaded outline maps available in field guides. All models were significantly better than random in both binomial tests of omission and receiver operating characteristic (ROC) analyses. The area under the ROC curve (AUC) was almost always higher for Maxent, indicating better discrimination of suitable versus unsuitable areas for the species. The Maxent modeling approach can be used in its present form for many applications with presence-only datasets, and merits further research and development.
Article
Broad-scale mapping of marine benthos is required for marine resource management and conservation. This study combines textural derivatives based on bathymetry from multibeam hydroacoustics with underwater video observations to model and map sessile biota between 10- and 60-m water depth over 35 km2 in Point Addis Marine National Park (MNP), Vic., Australia. Classification tree models and maps were developed for macroalgae (all types, mixed red algae, Ecklonia, and rhodoliths) and sessile invertebrates (all types, sponges, and ascidians). Model accuracy was tested on 25% of the video observation dataset reserved from modelling. Models fit well for most macroalgae categories (correct classification rates of 67–84%), but are not as good for sessile invertebrate classes (correct classification rates of 57–62%). The poor fit of the sessile invertebrate models may be the combined result of grouping organisms with different environmental requirements and the effect of false absences recorded during video interpretation due to poor image quality. Probability maps, binary single-class maps, and multi-class maps supply spatially explicit, detailed information on the distribution of sessile benthic biota within the MNP and provide information at a landscape-scale for ecological investigations and marine management.
Article
Distribution models are used increasingly for species conservation assessments over extensive areas, but the spatial resolution of the modeled data and, consequently, of the predictions generated directly from these models are usually too coarse for local conservation applications. Comprehensive distribution data at finer spatial resolution, however, require a level of sampling that is impractical for most species and regions. Models can be downscaled to predict distribution at finer resolutions, but this increases uncertainty because the predictive ability of models is not necessarily consistent beyond their original scale. We analyzed the performance of downscaled, previously published models of environmental favorability (a generalized linear modeling technique) for a restricted endemic insectivore, the Iberian desman (Galemys pyrenaicus), and a more widespread carnivore, the Eurasian otter (Lutra lutra), in the Iberian Peninsula. The models, built from presence-absence data at 10 x 10 km resolution, were extrapolated to a resolution 100 times finer (1 x 1 km). We compared downscaled predictions of environmental quality for the two species with published data on local observations and on important conservation sites proposed by experts. Predictions were significantly related to observed presence or absence of species and to expert selection of sampling sites and important conservation sites. Our results suggest the potential usefulness of downscaled projections of environmental quality as a proxy for expensive and time-consuming field studies when the field studies are not feasible. This method may be valid for other similar species if coarse-resolution distribution data are available to define high-quality areas at a scale that is practical for the application of concrete conservation measures.
Article
A new method is presented for flexible regression modeling of high dimensional data. The model takes the form of an expansion in product spline basis functions, where the number of basis functions as well as the parameters associated with each one (product degree and knot locations) are automatically determined by the data. This procedure is motivated by the recursive partitioning approach to regression and shares its attractive properties. Unlike recursive partitioning, however, this method produces continuous models with continuous derivatives. It has more power and flexibility to model relationships that are nearly additive or involve interactions in at most a few variables. In addition, the model can be represented in a form that separately identifies the additive contributions and those associated with the different multivariable interactions.