Comparison of alternative strategies for invasive species distribution modeling

Species distribution models (SDMs) can provide useful information for managing biological invasions, such as identification of priority areas for early detection or for determining containment boundaries. However, prediction of invasive species using SDMs can be challenging because they typically violate the core assumption of being at equilibrium with their environment, which may lead to poorly guided management resulting from high levels of omission. Our goal was to provide a suite of potential decision strategies (DSs) that were not reliant on the equilibrium assumption but rather could be chosen to better match the management application, which in this case was to ensure containment through adequate surveillance. We used presence-only data and expert knowledge for model calibration and presence/absence data to evaluate the potential distribution of an introduced mesquite (Leguminoseae: Prosopis) invasion located in the Pilbara Region of northwest Western Australia. Five different DSs with varying levels of conservatism/risk were derived from a multi-criteria evaluation model using ordered weighted averaging. The performance of DSs over all possible thresholds was examined using receiver operating characteristic (ROC) analysis. DSs not on the convex hull of the ROC curves were discarded. Two threshold determination methods (TDMs) were compared on the two remaining DSs, one that assumed equilibrium (by maximizing overall prediction success) and another that assumed the invasion was ongoing (using a 95% threshold for true positives). The most conservative DS fitted the validation data most closely but could only predict 75% of the presence data. A more risk-taking DS could predict 95% of the presence data, which identified 8.5 times more area for surveillance, and better highlighted known populations that are still rapidly invading. This DS and TDM coupling was considered to be the most appropriate for our management application. Our results show that predictive niche modeling was highly sensitive to risk levels, but that these can be tailored to match specified management objectives. The methods implemented can be readily adapted to other invasive species or for conservation purposes.     

Invasive species distribution modeling (iSDM): Are absence data and dispersal constraints needed to predict actual distributions?

Species distribution models (SDMs) based on statistical relationships between occurrence data and underlying environmental conditions are increasingly used to predict spatial patterns of biological invasions and prioritize locations for early detection and control of invasion outbreaks. However, invasive species distribution models (iSDMs) face special challenges because (i) they typically violate SDM's assumption that the organism is in equilibrium with its environment, and (ii) species absence data are often unavailable or believed to be too difficult to interpret. This often leads researchers to generate pseudo-absences for model training or utilize presence-only methods, and to confuse the distinction between predictions of potential vs. actual distribution. We examined the hypothesis that true-absence data, when accompanied by dispersal constraints, improve prediction accuracy and ecological understanding of iSDMs that aim to predict the actual distribution of biological invasions. We evaluated the impact of presence-only, true-absence and pseudo-absence data on model accuracy using an extensive dataset on the distribution of the invasive forest pathogen Phytophthora ramorum in California. Two traditional presence/absence models (generalized linear model and classification trees) and two alternative presence-only models (ecological niche factor analysis and maximum entropy) were developed based on 890 field plots of pathogen occurrence and several climatic, topographic, host vegetation and dispersal variables. The effects of all three possible types of occurrence data on model performance were evaluated with receiver operating characteristic (ROC) and omission/commission error rates. Results show that prediction of actual distribution was less accurate when we ignored true-absences and dispersal constraints. Presence-only models and models without dispersal information tended to over-predict the actual range of invasions. Models based on pseudo-absence data exhibited similar accuracies as presence-only models but produced spatially less feasible predictions. We suggest that true-absence data are a critical ingredient not only for accurate calibration but also for ecologically meaningful assessment of iSDMs that focus on predictions of actual distributions.     

Maximum entropy modeling of species geographic distributions

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.     

Spatial modeling in ecology: the flexibility of eigenfunction spatial analyses

Recently, analytical approaches based on the eigenfunctions of spatial configuration matrices have been proposed in order to consider explicitly spatial predictors. The present study demonstrates the usefulness of eigenfunctions in spatial modeling applied to ecological problems and shows equivalencies of and differences between the two current implementations of this methodology. The two approaches in this category are the distance-based (DB) eigenvector maps proposed by P. Legendre and his colleagues, and spatial filtering based upon geographic connectivity matrices (i.e., topology-based; CB) developed by D. A. Griffith and his colleagues. In both cases, the goal is to create spatial predictors that can be easily incorporated into conventional regression models. One important advantage of these two approaches over any other spatial approach is that they provide a flexible tool that allows the full range of general and generalized linear modeling theory to be applied to ecological and geographical problems in the presence of nonzero spatial autocorrelation.     

Investigating cumulative effects across ecological scales

Species, habitats, and ecosystems are increasingly exposed to multiple anthropogenic stressors, fueling a rapidly expanding research program to understand the cumulative impacts of these environmental modifications. Since the 1970s, a growing set of methods has been developed through two parallel, sometimes connected, streams of research within the applied and academic realms to assess cumulative effects. Past reviews of cumulative effects assessment (CEA) methods focused on approaches used by practitioners. Academic research has developed several distinct and novel approaches to conducting CEA. Understanding the suite of methods that exist will help practitioners and academics better address various ecological foci (physiological responses, population impacts, ecosystem impacts) and ecological complexities (synergistic effects, impacts across space and time). We reviewed 6 categories of methods (experimental, meta-analysis, single-species modeling, mapping, qualitative modeling, and multispecies modeling) and examined the ability of those methods to address different levels of complexity. We focused on research gaps and emerging priorities. We found that no single method assessed impacts across the 4 ecological foci and 6 ecological complexities considered. We propose that methods can be used in combination to improve understanding such that multimodel inference can provide a suite of comparable outputs, mapping methods can help prioritize localized models or experimental gaps, and future experiments can be paired from the outset with models they will inform.     

Modeling a spatially restricted distribution in the Neotropics: How the size of calibration area affects the performance of five presence-only methods

We here examine species distribution models for a Neotropical anuran restricted to ombrophilous areas in the Brazilian Atlantic Forest hotspot. We extend the known occurrence for the treefrog Hypsiboas bischoffi (Anura: Hylidae) through GPS field surveys and use five modeling methods (BIOCLIM, DOMAIN, OM-GARP, SVM, and MAXENT) and selected bioclimatic and topographic variables to model the species distribution. Models were first trained using two calibration areas: the Brazilian Atlantic Forest (BAF) and the whole of South America (SA). All modeling methods showed good levels of predictive power and accuracy with mean AUC ranging from 0.77 (BIOCLIM/BAF) to 0.99 (MAXENT/SA). MAXENT and SVM were the most accurate presence-only methods among those tested here. All but the SVM models calibrated with SA predicted larger distribution areas when compared to models calibrated in BAF. OM-GARP dramatically overpredicted the species distribution for the model calibrated in SA, with a predicted area around 10⁶ km² larger than predicted by other SDMs. With increased calibration area (and environmental space), OM-GARP predictions followed changes in the environmental space associated with the increased calibration area, while MAXENT models were more consistent across calibration areas. MAXENT was the only method that retrieved consistent predictions across calibration areas, while allowing for some overprediction, a result that may be relevant for modeling the distribution of other spatially restricted organisms.     

Incorporating anthropogenic variables into a species distribution model to map gypsy moth risk

This paper presents a novel methodology for multi-scale and multi-type spatial data integration in support of insect pest risk/vulnerability assessment in the contiguous United States. Probability of gypsy moth (Lymantria dispar L.) establishment is used as a case study. A neural network facilitates the integration of variables representing dynamic anthropogenic interaction and ecological characteristics. Neural network model (back-propagation network [BPN]) results are compared to logistic regression and multi-criteria evaluation via weighted linear combination, using the receiver operating characteristic area under the curve (AUC) and a simple threshold assessment. The BPN provided the most accurate infestation-forecast predictions producing an AUC of 0.93, followed by multi-criteria evaluation (AUC = 0.92) and logistic regression (AUC = 0.86) when independently validating using post model infestation data. Results suggest that BPN can provide valuable insight into factors contributing to introduction for invasive species whose propagation and establishment requirements are not fully understood. The integration of anthropogenic and ecological variables allowed production of an accurate risk model and provided insight into the impact of human activities.     

Experimental verification of ecological niche modeling in a heterogeneous environment

The current range of ecological habitats occupied by a species reflects a combination of the ecological tolerance of the species, dispersal limitation, and competition. Whether the current distribution of a species accurately reflects its niche has important consequences for the role of ecological niche modeling in predicting changes in species ranges as the result of biological invasions and climate change. We employed a detailed data set of species occurrence and spatial variation in biotic and abiotic attributes to model the niche of a native California annual plant, Collinsia sparsiflora. We tested the robustness of our model for both the realized and fundamental niche by planting seeds collected from four populations, representing two ecotypes, into plots that fully represented the five-dimensional niche space described by our model. The model successfully predicted which habitats allowed for C. sparsiflora persistence, but only for one of the two source ecotypes. Our results show that substantial niche divergence has occurred in our sample of four study populations, illustrating the importance of adequately sampling and describing within-species variation in niche modeling.     

Coexistence of the niche and neutral perspectives in community ecology

The neutral theory for community structure and biodiversity is dependent on the assumption that species are equivalent to each other in all important ecological respects. We explore what this concept of equivalence means in ecological communities, how such species may arise evolutionarily, and how the possibility of ecological equivalents relates to previous ideas about niche differentiation. We also show that the co-occurrence of ecologically similar or equivalent species is not incompatible with niche theory as has been supposed, because niche relations can sometimes favor coexistence of similar species. We argue that both evolutionary and ecological processes operate to promote the introduction and to sustain the persistence of ecologically similar and in many cases nearly equivalent species embedded in highly structured food webs. Future work should focus on synthesizing niche and neutral perspectives rather than dichotomously debating whether neutral or niche models provide better explanations for community structure and biodiversity.     

What does ecological modelling model? A proposed classification of ecological niche models based on their underlying methods

Species distribution model is the term most frequently used in ecological modelling, but other authors used instead predictive habitat distribution model or species-habitat models. A consensual ecological modelling terminology that avoids misunderstandings and takes into account the ecological niche theory does not exist at present. Moreover, different studies differ in the type of niche that is represented by similar distribution models. I propose to use as standard ecological modelling terminology the terms “ecological niche”, “potential niche”, “realized niche” models (for modelling their respective niches), and “habitat suitability map” (for the output of the niche models). Therefore, the user can understand more easily that models always forecast species’ niche and relate more closely the different types of niche models.     

Movement distances enhance validity of predictive models

Including the distance species are able to move in predictive models improves conservation practice. Bird inventory projects carried out from 1993 to 2004 in Taiwan provide an opportunity to investigate the relationships among species distribution, movement distance, and the environment. We compared projected distributions of 17 Taiwanese endemic bird species using what we called the Standard Method (i.e. movement distance is zero) and what we called the Buffer Method (i.e. movement distance is longer than zero) in three presence-only models (GARP, MAXENT and LIVES). The Standard Method used species original occurrence records directly while the Buffer Method expanded the occurrence of species to areas 1 km² around each recorded location. We first tested the efficacy of the Buffer Method using ten common species of the 17, and then applied the method to two rare species of the 17. For both the common and rare species, the distributions predicted by the two methods showed slight but important differences. The Buffer Method for all species had a higher average predictive probability, while the Standard Method had a higher maximum predictive probability. Most of the values for the area under the curve (AUC) were over 0.8 with the exceptions of Taiwan Barbet (Megalaima nuchalis) and Taiwan Hwamei (Garrulax taewanus), which have recently separated from Indochinese Barbet (Megalaima annamensis) and Chinese Hwamei (Garrulax canorus), and since 2008 and 2006 have been regarded as species endemic to the study area. Kappa values showed good performance for all species using both methods. The Buffer Method, however, resulted in significantly higher sensitivity and accuracy values for all models of species (p < 0.05). We conclude that when modeling species distribution including the area where the species was censused along with areas within the minimum movement areas better defines the surrounding areas that might supplement core habitat requirements. Therefore, using the Buffer Method, species surrounding distribution can be obtained which provides a better understanding of the species distributions. Given that distribution size is a key to the conservation of species, we suggest the Buffer Method can be used in conservation planning.     

Testing the Generality of Bird-Habitat Models

Bird-habitat models are frequently used as predictive modeling tools—for example, to predict how a species will respond to habitat modifications. We investigated the generality of the predictions from this type of model. Multivariate models were developed for Golden Eagle (Aquila chrysaetos), Raven (Corvus corax), and Buzzard (Buteo buteo) living in northwest Scotland. Data were obtained for all habitat and nest locations within an area of 2349 km². This assemblage of species is relatively static with respect to both occupancy and spatial positioning. The area was split into five geographic subregions: two on the mainland and three on the adjacent Island of Mull, which has one of United Kingdom's richest raptor fauna assemblages. Because data were collected for all nest locations and habitats, it was possible to build models that did not incorporate sampling error. A range of predictive models was developed using discriminant analysis and logistic regression. The models differed with respect to the geographical origin of the data used for model development. The predictive success of these models was then assessed by applying them to validation data. The models showed a wide range of predictive success, ranging from only 6% of nest sites correctly predicted to 100% correctly predicted. Model validation techniques were used to ensure that the models' predictions were not statistical artefacts. The variability in prediction success seemed to result from methodological and ecological processes, including the data recording scheme and interregional differences in nesting habitat. The results from this study suggest that conservation biologists must be very careful about making predictions from such studies because we may be working with systems that are inherently unpredictable.     

The models of niche and their application

Having studied the definitions of niche proposed by different ecologists, I have proposed a quantitative method of niche which can be applied to plants. Accordingly, the niche of an operational taxonomic unit (OTU) has been described by a mapping from its environmental set to the unit interval [0, 1], which enables a model of niche to be mathematically operational. The concepts of fundamental niche, realized niche, time niche, etc., may be described by using mathematical models related to each other, and the geometrical relationships between them can be revealed by a multi-dimensional surface. The uni-factor models are built upon the condition that the other factors are optimal for OTU, which are particular cases of the multi-factor models. The establishment of the quantitative relationship between these two kinds of models makes it possible to find out the plants' fundamental niche by doing uni-factor experiments. This may simplify the experiments in which the parameters in a practically applied model are to be estimated. The niche index introduced in this paper is related to average level and aftereffect of plant responses to the effects of its environment (i.e. “inertia”), thus it should be the basis of the simulation of plant seed yield and of its environmental evaluation. Accordingly, models of niche index, of plant seed yield, of plant growth and of environmental evaluation have been built which can be applied to the environmental evaluation or the prediction and management of plant (crop) production, etc.As an example of application, the models of wheat yield and its environmental assessment have been established and practically tested. The results of testing the model of wheat yield showed that the relative errors are 8% and 7.2%, respectively, in 1984 and 1986. The results of the environmental assessment of wheat reveal the fact that the insufficiency of the soil moisture at the 2th and 3th stages is the main restriction of the production of the wheat in Dinxi, Gansu Province, China.     

The Use of Bayesian Model Averaging to Better Represent Uncertainty in Ecological Models

Abstract:   In conservation biology, uncertainty about the choice of a statistical model is rarely considered. Model-selection uncertainty occurs whenever one model is chosen over plausible alternative models to represent understanding about a process and to make predictions about future observations. The standard approach to representing prediction uncertainty involves the calculation of prediction (or confidence) intervals that incorporate uncertainty about parameter estimates contingent on the choice of a "best" model chosen to represent truth. However, this approach to prediction based on statistical models tends to ignore model-selection uncertainty, resulting in overconfident predictions. Bayesian model averaging (BMA) has been promoted in a range of disciplines as a simple means of incorporating model-selection uncertainty into statistical inference and prediction. Bayesian model averaging also provides a formal framework for incorporating prior knowledge about the process being modeled. We provide an example of the application of BMA in modeling and predicting the spatial distribution of an arboreal marsupial in the Eden region of southeastern Australia. Other approaches to estimating prediction uncertainty are discussed.     

Spatial and temporal hydrodynamic and water quality modeling analysis of a large reservoir on the South Carolina (USA) coastal plain

Two-dimensional, 31-segment, 61-channel hydrodynamic and water quality models of Lake Marion (surface area 330.7 km²; volume 1548.3×10⁶ m³) were developed using the WASP5 modeling system. Field data from 1985 to 1990 were used to parameterize the models. Phytoplankton kinetic rates and constants were obtained from a related in situ study; others from modeling literature. The hydrodynamic model was calibrated to estimates of daily lake volume; the water quality model was calibrated for ammonia, nitrate, ortho-phosphate, dissolved oxygen, chlorophyll-a, biochemical oxygen demand, organic nitrogen, and organic phosphorus. Water quality calibration suggested the model characterized phytoplankton and nutrient dynamics quite well. The model was validated (Kolmogorov–Smirnov two-sample goodness-of-fit test at P<0.05) by reparameterizing the nutrient loading functions using an independent set of field data. The models identified several factors that may contribute to the spatial variability previously reported from other research in the reservoir, despite the superficial absence of complex structure. Sensitivity analysis of the phytoplankton kinetic rates suggest that study site-specific estimates were important for obtaining model fit to field data. Sediment sources of ammonia (10–60 mg m⁻² day⁻¹) and phosphate (1–6 mg m⁻² day⁻¹) were important to achieve model calibration, especially during periods of high temperatures and low dissolved oxygen. This sediment flux accounted for 78% (nitrogen) and 50% (phosphorus) of the annual load. Spatial and temporal variability in the lake, reflected in the calibrated and validated models, suggest that ecological factors that influence phytoplankton productivity and nutrient dynamics are different in various parts of the lake. The WASP5 model as implemented here does not fully accommodate the ecological variability in Lake Marion due to model constraints on the specification of rate constants. This level of spatial detail may not be appropriate for an operational reservoir model, but as a research tool the models are both versatile and useful.     

An evaluation of environmental factors affecting species distributions

Many different models can be built to explain the distributions of species. Often there is no single model that is clearly better than the alternatives, and this leads to uncertainty over which environmental factors are limiting species’ distributions. We investigated the support for different environmental factors by determining the drop in model performance when selected predictors were excluded from the model building process. We used a paired t-test over 37 plant species so that an environmental factor was only deemed significant if it consistently improved the results for multiple species. Geology and winter minimum temperatures were found to be the environmental factors with the most support, with a significant drop in model performance when either of these factors was excluded. However, there was less support for summer maximum temperature, as other environmental factors could combine to produce similar model performance. Our method of evaluating environmental factors using multiple species will not be capable of detecting predictors that are only important for one or two species, but it is difficult to distinguish these from spurious correlations. The strength of the method is that it increases inference for factors that consistently affect the distributions of many species. We discourage the assessment of models against predefined benchmarks, such as an area under the curve (AUC) of more than 0.7, as many alternative models for the same species produce similar results. Therefore, the benchmarks do not provide any indication of how the performance of the selected model compares to alternative models, and they provide weak inference to accept any selected model.     

Cross-validation of species distribution models: removing spatial sorting bias and calibration with a null model

Species distribution models are usually evaluated with cross-validation. In this procedure evaluation statistics are computed from model predictions for sites of presence and absence that were not used to train (fit) the model. Using data for 226 species, from six regions, and two species distribution modeling algorithms (Bioclim and MaxEnt), I show that this procedure is highly sensitive to "spatial sorting bias": the difference between the geographic distance from testing-presence to training-presence sites and the geographic distance from testing-absence (or testing-background) to training-presence sites. I propose the use of pairwise distance sampling to remove this bias, and the use of a null model that only considers the geographic distance to training sites to calibrate cross-validation results for remaining bias. Model evaluation results (AUC) were strongly inflated: the null model performed better than MaxEnt for 45% and better than Bioclim for 67% of the species. Spatial sorting bias and area under the receiver-operator curve (AUC) values increased when using partitioned presence data and random-absence data instead of independently obtained presence-absence testing data from systematic surveys. Pairwise distance sampling removed spatial sorting bias, yielding null models with an AUC close to 0.5, such that AUC was the same as null model calibrated AUC (cAUC). This adjustment strongly decreased AUC values and changed the ranking among species. Cross-validation results for different species are only comparable after removal of spatial sorting bias and/or calibration with an appropriate null model.     

Habitat classification modeling with incomplete data: pushing the habitat envelope.

Habitat classification models (HCMs) are invaluable tools for species conservation, land-use planning, reserve design, and metapopulation assessments, particularly at broad spatial scales. However, species occurrence data are often lacking and typically limited to presence points at broad scales. This lack of absence data precludes the use of many statistical techniques for HCMs. One option is to generate pseudo-absence points so that the many available statistical modeling tools can bb used. Traditional techniques generate pseudo-absence points at random across broadly defined species ranges, often failing to include biological knowledge concerning the species-habitat relationship. We incorporated biological knowledge of the species-habitat relationship into pseudo-absence points by creating habitat envelopes that constrain the region from which points were randomly selected. We define a habitat envelope as an ecological representation of a species, or species feature's (e.g., nest) observed distribution (i.e., realized niche) based on a single attribute, or the spatial intersection of multiple attributes. We created HCMs for Northern Goshawk (Accipiter gentilis atricapillus) nest habitat during the breeding season across Utah forests with extant nest presence points and ecologically based pseudo-absence points using logistic regression. Predictor variables were derived from 30-m USDA Landfire and 250-m Forest Inventory and Analysis (FIA) map products. These habitat-envelope-based models were then compared to null envelope models which use traditional practices for generating pseudo-absences. Models were assessed for fit and predictive capability using metrics such as kappa, threshold-independent receiver operating characteristic (ROC) plots, adjusted deviance (D(adj)2), and cross-validation, and were also assessed for ecological relevance. For all cases, habitat envelope-based models outperformed null envelope models and were more ecologically relevant, suggesting that incorporating biological knowledge into pseudo-absence point generation is a powerful tool for species habitat assessments. Furthermore, given some a priori knowledge of the species-habitat relationship, ecologically based pseudo-absence points can be applied to any species, ecosystem, data resolution, and spatial extent.     

Comparison of Two Types of Metapopulation Models in Real and Artificial Landscapes

Abstract: Application of metapopulation models is becoming increasingly widespread in the conservation of species in fragmented landscapes. We provide one of the first detailed comparisons of two of the most common modeling techniques, incidence function models and stage-based matrix models, and test their accuracy in predicting patch occupancy for a real metapopulation. We measured patch occupancies and demographic rates for regional populations of the Florida scrub lizard (   Sceloporus woodi ) and compared the observed occupancies with those predicted by each model. Both modeling strategies predicted patch occupancies with good accuracy ( 77–80%) and gave similar results when we compared hypothetical management scenarios involving removal of key habitat patches and degradation of habitat quality. To compare the two modeling approaches over a broader set of conditions, we simulated metapopulation dynamics for 150 artificial landscapes composed of equal-sized patches (2–1024 ha) spaced at equal distances (50–750 m). Differences in predicted patch occupancy were small to moderate (<20%) for about 74% of all simulations, but 22% of the landscapes had differences openface> 50%. Incidence function models and stage-based matrix models differ in their approaches, assumptions, and requirements for empirical data, and our findings provide evidence that the two models can produce different results. We encourage researchers to use both techniques and further examine potential differences in model output. The feasibility of obtaining data for population modeling varies widely among species and limits the modeling approaches appropriate for each species. Understanding different modeling approaches will become increasingly important as conservation programs undertake the challenge of managing for multiple species in a landscape context.     

Support vector machines for predicting distribution of Sudden Oak Death in California

In the central California coastal forests, a newly discovered virulent pathogen (Phytophthora ramorum) has killed hundreds of thousands of native oak trees. Predicting the potential distribution of the disease in California remains an urgent demand of regulators and scientists. Most methods used to map potential ranges of species (e.g. multivariate or logistic regression) require both presence and absence data, the latter of which are not always feasibly collected, and thus the methods often require the generation of ‘pseudo’ absence data. Other methods (e.g. BIOCLIM and DOMAIN) seek to model the presence-only data directly. In this study, we present alternative methods to conventional approaches to modeling by developing support vector machines (SVMs), which are the new generation of machine learning algorithms used to find optimal separability between classes within datasets, to predict the potential distribution of Sudden Oak Death in California. We compared the performances of two types of SVMs models: two-class SVMs with ‘pseudo’ absence data and one-class SVMs. Both models performed well. The one-class SVMs have a slightly better true-positive rate (0.9272 ± 0.0460 S.D.) than the two-class SVMs (0.9105 ± 0.0712 S.D.). However, the area predicted to be at risk for the disease using the one-class SVMs (18,441 km²) is much larger than that of the two-class SVMs (13,828 km²). Both models show that the majority of disease risk will occur in coastal areas. Compared with the results of two-class SVMs, the one-class SVMs predict a potential risk in the foothills of the Sierra Nevada mountain ranges; much greater risks are also found in Los Angles and Humboldt Counties. We believe the support vector machines when coupled with geographic information system (GIS) will be a useful method to deal with presence-only data in ecological analysis over a range of scales.