MaxEnt Predictions Research Articles

Comparison of two maximum entropy models highlights the metabolic structure of metacommunities as a key determinant of local community assembly

The principle of Maximum Entropy (MaxEnt) promises a novel approach for understanding community assembly. Despite reproducing a variety of observed species abundance patterns, MaxEnt models in ecology have been hampered by disparate model assumptions and interpretations. A recurring challenge is that MaxEnt predictions are highly sensitive to the level of detail used to describe the community being modeled, and there seems to be no reason to prefer one level of detail over another. Here we present of formal unification of two previously developed MaxEnt models which differ in their level of detail, but which are otherwise mathematically similar. The less detailed model, “Maximum Entropy Theory of Ecology” (METE), does not resolve species identity or explicitly represent species-specific traits. The more detailed model, “Very Entropic Growth” (VEG), defines each separate species by its per capita metabolic rate ε and assumes a “density of species” function ρ(ε) representing the distribution of ε in the metacommunity. A formal comparison of METE and VEG then highlights ρ(ε) as a key determinant of local community assembly. In particular, appropriate choice of ρ(ε) in VEG can produce more realistic predictions for the metabolic-rank distribution of local communities than METE, which does not explicitly account for metacommunity structure. This opens new avenues of inquiry about what determines metacommunity structure in nature and suggests possible ways to improve METE.

Prediction of Large Whale Distributions: A Comparison of Presence–Absence and Presence-Only Modeling Techniques

Species distribution models that predict species occurrence or density by quantifying relationships with environmental variables are used for a variety of scientific investigations and management applications. For endangered species, such as large whales, models help to understand the ecological factors influencing variability in distributions and to assess potential risk from shipping, fishing, and other human activities. Systematic surveys record species presence and absence, as well as the associated search effort, but are very expensive. Presence-only data consisting only of sightings can increase sample size, but may be biased in both geographical and niche space. We built generalized additive models (GAM) using presence-absence sightings data and maximum entropy models (Maxent) using the same presence-absence sightings data, and also using presence-only sightings data, for four large whale species in the eastern tropical Pacific Ocean: humpback (Megaptera novaeangliae), blue (Balaenoptera musculus), Bryde’s (Balaenoptera edeni), and sperm whales (Physeter macrocephalus). Environmental variables were surface temperature, surface salinity, thermocline depth, stratification index, and seafloor depth. We compared predicted distributions and environmental response curves from each of the two model types. Maxent and GAM model predictions based on systematic survey data are very similar, when Maxent absences are selected from the survey trackline data. However, we show that spatial bias in presence-only Maxent predictions can be caused by using pseudo-absences instead of observed absences and by the sampling biases of both opportunistic data and stratified systematic survey data with uneven coverage between strata. Predictions of uncommon large whale distributions from Maxent or other presence-only techniques may be useful for science or management, but only if spatial bias in the observations is addressed in the derivation and interpretation of model predictions.

Ecology suitability study of Trollius chinensis

Based on the distribution information of 110 samples and 55 environmental factors, Maxent model was used to predict the ecology suitability regions of Trollius chinensis. The study aims at providing theory basis for the cultivation of T. chinensis. The results showed that the Maxent model prediction result was good (AUC>0.9) and the main factors effecting the ecology suitability regions of T. chinensis were precipitation in July, standard deviation of seasonal variation of temperature, annual mean temperature, precipitation in August and altitude. The ecology suitable regions of T. chinensis mainly concentrated in Shanxi, Hebei, east of Inner Mongolia, west of Jilin and Liaoning, north of Shaanxi, south of Ningxia, east and south of Gansu, and east of Qinghai. The results indicated that except for traditional distribution regions, north of Shaanxi, south of Ningxia, east and south of Gansu, and east of Qinghai could selected as the regions for cultivation of T. chinensis. It provides theory basis for selecting suitable regions to cultivate T. chinensis.

Inferring climatic controls of rice stem borers’ spatial distributions using maximum entropy modelling

AbstractAccurate assessment of pest potential distributions is needed to identify their establishment risks that play a key role in pest management in agricultural ecosystems. We used a correlative niche modelling method (Maxent) to predict and map the spatial distributions of two important rice stem borers, Chilo suppressalis and Sesamia cretica, in paddy fields of Iran. In total, 195 presence occurrence records (101 records for C. suppressalis and 94 records for S. cretica) were compiled. A set of environmental and topographic variables, with the highest effects on the species distributions and the lowest correlations among themselves, were used. The results showed that mainly the northern parts of Iran were the most suitable areas for C. suppressalis, and north, north‐east and south‐west of Iran as the most suitable areas for S. cretica. Both models performed well, with an area under the receiver operating characteristic curve (AUC) of 0.983 and 0.786 for C. suppressalis and S. cretica, respectively. The Maxent models showed higher accuracy for predicting the distribution of the specialist pest with the small range sizes compared to the generalist species. Assessing the importance of environmental variables, which were derived from the jackknife test, showed the precipitation as the variable with the highest contribution (66%) in explaining the spatial distribution of C. suppressalis compared to the other variables. The distribution of S. cretica was influenced by a set of variables derived from both the precipitation and temperature. The Maxent predictions were useful to map the geographical distributions of the risk for both rice stem borers that is needed to develop effective management strategies.

Species Distribution Modelling: Contrasting presence-only models with plot abundance data.

Species distribution models (SDMs) are widely used in ecology and conservation. Presence-only SDMs such as MaxEnt frequently use natural history collections (NHCs) as occurrence data, given their huge numbers and accessibility. NHCs are often spatially biased which may generate inaccuracies in SDMs. Here, we test how the distribution of NHCs and MaxEnt predictions relates to a spatial abundance model, based on a large plot dataset for Amazonian tree species, using inverse distance weighting (IDW). We also propose a new pipeline to deal with inconsistencies in NHCs and to limit the area of occupancy of the species. We found a significant but weak positive relationship between the distribution of NHCs and IDW for 66% of the species. The relationship between SDMs and IDW was also significant but weakly positive for 95% of the species, and sensitivity for both analyses was high. Furthermore, the pipeline removed half of the NHCs records. Presence-only SDM applications should consider this limitation, especially for large biodiversity assessments projects, when they are automatically generated without subsequent checking. Our pipeline provides a conservative estimate of a species’ area of occupancy, within an area slightly larger than its extent of occurrence, compatible to e.g. IUCN red list assessments.

Quantifying apart what belongs together: A multi‐state species distribution modelling framework for species using distinct habitats

Abstract Species distribution models (SDMs) have been used to inform scientists and conservationists about the status and change in occurrence patterns in threatened species. Many mobile species use multiple functionally distinct habitats, and cannot occupy one habitat type without the other being within a reachable distance. For such species, classical applications of SDMs might lead to erroneous representations of habitat suitability, as the complex relationships between predictors are lost when merging occurrence information across multiple habitats. To better account for the spatial arrangement of complementary—yet mandatory—habitat types, it is important to implement modelling strategies that partition occurrence information according to habitat use in a spatial context. Here, we address this issue by introducing a multi‐state SDM framework. The multi‐state SDM framework stratifies occurrences according to the temporal or behavioural use of distinct habitat types, referred to as “states.” Multiple SDMs are then run for each state and statistical thresholds of presence are used to combine these separate predictions. To identify suitable sites that account for distance between habitats, two optional modules are proposed where the thresholded output is aggregated and filtered by minimum area size, or through moving windows across maximum reachable distances. We illustrate the full use of this framework by modelling the dynamic terrestrial breeding habitat preferences of the New Zealand sea lion (NZSL) (Phocarctos hookeri), using Maxent and trialling both modules to identify suitable sites for possible recolonization. The Maxent predictions showed excellent performance, and the multi‐state SDM framework highlighted 36–77 potential suitable breeding sites in the study area. This framework can be applied to inform management when defining habitat suitability for species with complex changes in habitat use. It accounts for temporal and behavioural changes in distribution, maintains the individuality of each partitioned SDM, and considers distance between distinct habitat types. It also yields one final, easy‐to‐understand output for stakeholders and managers.

Factors restricting the range expansion of the invasive green anole Anolis carolinensis on Okinawa Island, Japan.

The green anole Anolis carolinensis invaded the Ogasawara Islands in Japan, drove various native species to extinction, and its distribution expanded 14 years after initial establishment. A. carolinensis invaded Okinawa Island, but it has not expanded its distribution in more than 25 years, although its density is extremely high in the southern region. To determine whether A. carolinensis has the potential to expand its distribution on Okinawa Island, we performed phylogenetic analysis of mitochondrial ND2 DNA sequences to study the origin of A. carolinensis that invaded Okinawa Island. We further used a species distribution model (MaxEnt) based on the distribution of native populations in North America to identify ecologically suitable areas on Okinawa Island. Nucleotide sequence analysis shows that the invader A. carolinensis originated in the western part of the Gulf Coast and inland areas of the United States and that a portion of the anoles on Okinawa was not introduced via the Ogasawara Islands. The MaxEnt predictions indicate that most areas in Okinawa Island are suitable for A. carolinensis. Therefore, A. carolinensis may have the potential to expand its distribution in Okinawa Island. The predictions indicate that habitat suitability is high in areas of high annual mean temperature and urbanized areas. The values of precipitation in summer in the northern region of Okinawa Island were higher compared with those of North America, which reduced the habitat suitability in Okinawa Island. Adaptation to low temperatures, an increase in the mean temperature through global warming, and an increase in open environments through land development will likely expand the distribution of A. carolinensis in Okinawa Island. Therefore, we must continue to monitor the introduced populations and be alert to the possibility that city planning that increases open environments may cause their range to expand.

Identification of High-Risk Agro-Ecological Regions using Species Distribution Modeling of Priority Invasive Species in Sri Lanka

Mimosa pigra, Annona glabra, Lantana camara, Prosopis juliflora and Parthenium histerophorus are considered as priority invasive plants of Sri Lanka that cause considerable impact on agriculture and food security. The potential areas of these species are not available for land managers for timely control and management measures. This study modelled the suitable climate area for these invasive species under current climate scenarios using Maxent model. The resulting models were found to have good discrimination ability of presence and absence, with very high AUC values. The study identified highly responding variables for the model prediction of each species. Maxent predictions were overlaid on agro-ecological map of Sri Lanka and high-risk agro-ecological regions were identified for the five species considered. Resulting maps can be a vital tool for land managers to take information based decisions to control and manage invasive species.

Ecology suitability study of Chinese materia medica Gentianae Macrophyllae Radix

This paper is aimed to predict ecology suitability distribution of Gentianae Macrophyllae Radix and search the main ecological factors affecting the suitability distribution. The 313 distribution information about G. macrophylla, 186 distribution information about G. straminea, 343 distribution information about G. dauricaand 131 distribution information about G. crasicaulis were collected though investigation and network sharing platform data . The ecology suitable distribution factors for production Gentianae Macrophyllae Radix was analyzed respectively by the software of ArcGIS and MaxEnt with 55 environmental factors. The result of MaxEnt prediction was very well (AUC was above 0.9). The results of predominant factors analysis showed that precipitation and altitude were all the major factors impacting the ecology suitable of Getiana Macrophylla Radix production. G. macrophylla ecology suitable region was mainly concentrated in south of Gansu, Shanxi, central of Shaanxi and east of Qinghai provinces. G. straminea ecology suitable region was mainly concentrated in southwest of Gansu, east of Qinghai, north and northwest of Sichuan, east of Xizang province. G. daurica ecology suitable region was mainly concentrated in south and southwest of Gansu, east of Qinghai, Shanxi and north of Shaanxi province. G. crasicaulis ecology suitable region was mainly concentrated in Sichuan and north of Yunnan, east of Xizang, south of Gansu and east of Qinghai province. The ecological suitability distribution result of Gentianae Macrophyllae Radix was consistent with each species actual distribution. The study could provide reference for the collection and protection of wild resources, meanwhile, provide the basis for the selection of cultivation area of Gentiana Macrophylla Radix.

Comparing Suitable Habitat Models to Predict Rare and Endemic Plant Species Distributions: What are the Limits of the Niche of Cola lorougnonis (Malvaceae) in Cote d'Ivoire?

<p class="1Body"><em>Cola lorougnonis</em> is an endemic, rare, and treated species. It was recently recorded in Dekpa forest near Agbaou (a village of Divo region, Cote d’Ivoire). In the same forest, 20 other rare and endemic plant species were recorded. Accurate modeling of geographical distributions of these species is crucial to various applications in ecology and biodiversity conservation. The present study analyzed suitable habitat models<strong> </strong>for the 21 species. The main objective was to test geographical predictions for the focused species: <em>Cola lorougnonis</em>. We used Maxent modelling method for predicting potential suitable habitats combining environmental variables and species records. We evaluated Maxent predictions using the area under the receiver-operating characteristic curve (AUC). For each species, the map of distribution was engineered using DIVA-GIS. We compared the suitable habitat areas among species. Principal Canonical Analysis allowed the ordination of species according to environmental variables. AUC values allowed to get 11 species with excellent distribution models, 8 species with good distribution models, and 2 species with predictive models considered as acceptable. <em>Cola lorougnonis</em> (AUC = 0.99) and <em>Drypetes singroboensis</em> (AUC = 0.96) have the same focused area: moist semi-deciduous forest in Cote d’Ivoire. They are more sensitive to changes in rainfall of both warmest and coldest seasons.<strong> </strong>The State of Côte d’Ivoire has to undertake monitoring, assessment and reporting of conservation status facilitation for all habitats where these species could be found within the territory. Additional studies focusing on the investigation areas and niche models onto future conditions of climate could be considered.</p>

Population Genetic Structure of the Asian Black Bear (Ursus thibetanus) within and Across Management Units in Northern Japan

Abstract. We examined the population genetic structure of the Asian black bear (Ursus thibetanus) in a large continuous habitat in northern Japan. To determine how population subdivision relates to management units (MUs) proposed by the Ministry of the Environment, genetic variation in the mitochondrial DNA control region (mtDNA-CR) and seven autosomal microsatellite loci was assessed in bears captured in southern Tohoku. Geographical distribution of the subpopulations was assessed using landscape analyses to find the best-fit model based on maximum entropy (MaxEnt) prediction and cost of movement. Genetic differentiation analyses revealed unique mtDNA-CR haplotypes in the study population. Bayesian population assignment of autosomal loci inferred four genetically distinct subpopulations in the study area (pop 1–4). The main distribution areas of pop1 and pop4 were separated by a high-cost zone for bear movement as a result of land use. Another high-cost zone was identified between pop2 and pop4, while th...

How climate change might influence the potential distribution of weed, bushmint (Hyptis suaveolens)?

Invasive species and climate change are considered as the most serious global environmental threats. In this study, we investigated the influence of projected global climate change on the potential distribution of one of the world's most successful invader weed, bushmint (Hyptis suaveolens (L.) Poit.). We used spatial data on 20 environmental variables at a grid resolution of 5km, and 564 presence records of bushmint from its native and introduced range. The climatic profiles ofthe native and invaded sites were analyzed in a multi-variate space in order to examine the differences in the position of climatic niches. Maximum Entropy (MaxEnt) model was used to predict the potential distribution of bushmint using presence records from entire range (invaded and native) along with 14 eco-physiologically relevant predictor variables. Subsequently, the trained MaxEnt model was fed with Hadley Centre Coupled Model (HadCM3) climate projections to predict potential distribution of bushmint by the year 2050 under A2a and B2a emission scenarios. MaxEnt predictions were very accurate with an Area Under Curve (AUC) value of 0.95. The results of Principal Component Analysis (PCA) indicated that climatic niche of bushminton the invaded sites is not entirely similar to its climatic niche in the native range. A vast area spread between 34° 02' north and 28° 18' south latitudes in tropics was predicted climatically suitable for bushmint. West and middle Africa, tropical southeast Asia, and northern Australia were predicted at high invasion risk. Study indicates enlargement, retreat, or shift across bushmint's invasion range under the influence of climate change. Globally, bushmint's potential distribution might shrink in future with more shrinkage for A2a scenario than B2a. The study outcome has immense potential for undertaking effective preventive/control measures and long-term management strategies forregions/countries, which are at higher risk of bushmint's invasion.

Climate change may alter breeding ground distributions of eastern migratory monarchs (Danaus plexippus) via range expansion of Asclepias host plants.

Climate change can profoundly alter species’ distributions due to changes in temperature, precipitation, or seasonality. Migratory monarch butterflies (Danaus plexippus) may be particularly susceptible to climate-driven changes in host plant abundance or reduced overwintering habitat. For example, climate change may significantly reduce the availability of overwintering habitat by restricting the amount of area with suitable microclimate conditions. However, potential effects of climate change on monarch northward migrations remain largely unknown, particularly with respect to their milkweed (Asclepias spp.) host plants. Given that monarchs largely depend on the genus Asclepias as larval host plants, the effects of climate change on monarch northward migrations will most likely be mediated by climate change effects on Asclepias. Here, I used MaxEnt species distribution modeling to assess potential changes in Asclepias and monarch distributions under moderate and severe climate change scenarios. First, Asclepias distributions were projected to extend northward throughout much of Canada despite considerable variability in the environmental drivers of each individual species. Second, Asclepias distributions were an important predictor of current monarch distributions, indicating that monarchs may be constrained as much by the availability of Asclepias host plants as environmental variables per se. Accordingly, modeling future distributions of monarchs, and indeed any tightly coupled plant-insect system, should incorporate the effects of climate change on host plant distributions. Finally, MaxEnt predictions of Asclepias and monarch distributions were remarkably consistent among general circulation models. Nearly all models predicted that the current monarch summer breeding range will become slightly less suitable for Asclepias and monarchs in the future. Asclepias, and consequently monarchs, should therefore undergo expanded northern range limits in summer months while encountering reduced habitat suitability throughout the northern migration.

Spatial extent of biotic interactions affects species distribution and abundance in river networks: the freshwater pearl mussel and its hosts

AbstractAimAt what spatial extent are biotic interactions discernible influences on the distribution and abundance of species in river networks? We address this question with analyses of data from river networks for Margaritifera margaritifera, a freshwater mussel that passes its larval stage attached to a host fish.LocationTwenty river networks in Galicia, north‐western Spain.MethodsA maximum‐entropy approach was implemented to model the species’ distribution. Geostatistical mixed models were used to analyse the mussel's abundance in dendritic river networks. Predictor variables included the abundance and biomass of host fish (biotic interactions) and abiotic predictors for climate, geology and land‐form.ResultsMaxent models of species distribution were improved by 4.5% in terms of the area under the receiver‐operating characteristic curve (AUC) by the inclusion of biotic interactions. Host‐fish predictors contributed 63% of the Maxent model prediction of mussel presence. A geostatistical mixed model explained 52% of the variance in mussel abundance when including all the mussel abundance sites in the study area; abiotic predictors had no significance and salmonid biomass and resident trout population density were the only significant biotic predictors, together explaining 2.4% of the variance. An autocovariate representing biotic interactions between mussels and fish explained 11.7% of the variance. Using only sites where migratory host fish were present (n = 149), a mixed model explained more variance (78%) and the contribution from the autocovariate for parasite–host interactions was about three times larger than for the model including all sites (n = 419). The spatial autocorrelation from mussel–fish interactions had a spatial extent (geostatistical range) greater than 15 km.Main conclusionsInteractions between mussels and their larval hosts in river networks are manifested in spatial patterns of species distribution and abundance in this region, encompassing 20 river networks. The directional topology of dendritic river networks strongly supports the upstream dispersal of mussels by parasitized host fish as a component of spatial autocorrelation in mussel abundance.

Variational principle underlying scale invariant social systems

MaxEnt's variational principle, in conjunction with Shannon's logarithmic information measure, yields only exponential functional forms in straightforward fashion. In this communication we show how to overcome this limitation via the incorporation, into the variational process, of suitable dynamical information. As a consequence, we are able to formulate a somewhat generalized Shannonian Maximum Entropy approach which provides a unifying "thermodynamic-like" explanation for the scale-invariant phenomena observed in social contexts, as city-population distributions. We confirm the MaxEnt predictions by means of numerical experiments with random walkers, and compare them with some empirical data.

Modeling the habitat suitability for deep-water gorgonian corals based on terrain variables

The coral species Paragorgia arborea and Primnoa resedaeformis are abundant and widely distributed gorgonians in North Atlantic waters. Both species add significant habitat complexity to the benthic environment, and support a host of invertebrate species. Mapping their distribution is an essential step in conservation and resource management, but challenging as a result of their remoteness. In this study, three predictive models — Ecological Niche Factor Analysis, Genetic Algorithm for Rule-set Production and Maximum Entropy modeling (MaxEnt) were applied to predict the distribution of species' suitable habitat across a region of Røst Reef (Norwegian margin) based on multiscale terrain variables.All three models were successful in predicting the habitat suitability for both gorgonian species across the study area, and the MaxEnt predictions were shown to outperform other predictions. All three models predicted the most suitable habitats for both species to mainly occur along the ridges and on the upper section of the large slide, suggesting both species preferentially colonize topographic highs. Jackknife tests for MaxEnt predictions highlighted the seabed aspect in relation to P. arborea distribution, and the seabed relative position (curvature) in relation to the distribution of both species. Given the vulnerability of deep-water corals to anthropogenic impacts, further comparative study over a wider study area would be particularly beneficial for the management of the species.

Likelihood analysis of species occurrence probability from presence‐only data for modelling species distributions

Summary1. Understanding the factors affecting species occurrence is a pre‐eminent focus of applied ecological research. However, direct information about species occurrence is lacking for many species. Instead, researchers sometimes have to rely on so‐called presence‐only data (i.e. when no direct information about absences is available), which often results from opportunistic, unstructured sampling. maxent is a widely used software program designed to model and map species distribution using presence‐only data.2. We provide a critical review of maxent as applied to species distribution modelling and discuss how it can lead to inferential errors. A chief concern is that maxent produces a number of poorly defined indices that are not directly related to the actual parameter of interest – the probability of occurrence (ψ). This focus on an index was motivated by the belief that it is not possible to estimate ψ from presence‐only data; however, we demonstrate that ψ is identifiable using conventional likelihood methods under the assumptions of random sampling and constant probability of species detection.3. The model is implemented in a convenient r package which we use to apply the model to simulated data and data from the North American Breeding Bird Survey. We demonstrate that maxent produces extreme under‐predictions when compared to estimates produced by logistic regression which uses the full (presence/absence) data set. We note that maxent predictions are extremely sensitive to specification of the background prevalence, which is not objectively estimated using the maxent method.4. As with maxent, formal model‐based inference requires a random sample of presence locations. Many presence‐only data sets, such as those based on museum records and herbarium collections, may not satisfy this assumption. However, when sampling is random, we believe that inference should be based on formal methods that facilitate inference about interpretable ecological quantities instead of vaguely defined indices.

The performance of range maps and species distribution models representing the geographic variation of species richness at different resolutions

ABSTRACTAim The method used to generate hypotheses about species distributions, in addition to spatial scale, may affect the biodiversity patterns that are then observed. We compared the performance of range maps and MaxEnt species distribution models at different spatial resolutions by examining the degree of similarity between predicted species richness and composition against observed values from well‐surveyed cells (WSCs).Location Mexico.Methods We estimated amphibian richness distributions at five spatial resolutions (from 0.083° to 2°) by overlaying 370 individual range maps or MaxEnt predictions, comparing the similarity of the spatial patterns and correlating predicted values with the observed values for WSCs. Additionally, we looked at species composition and assessed commission and omission errors associated with each method.Results MaxEnt predictions reveal greater geographic differences in richness between species rich and species poor regions than the range maps did at the five resolutions assessed. Correlations between species richness values estimated by either of the two procedures and the observed values from the WSCs increased with decreasing resolution. The slopes of the regressions between the predicted and observed values indicate that MaxEnt overpredicts observed species richness at all of the resolutions used, while range maps underpredict them, except at the finest resolution. Prediction errors did not vary significantly between methods at any resolution and tended to decrease with decreasing resolution. The accuracy of both procedures was clearly different when commission and omission errors were examined separately.Main conclusions Despite the congruent increase in the geographic richness patterns obtained from both procedures as resolution decreases, the maps created with these methods cannot be used interchangeably because of notable differences in the species compositions they report.

Maximum Entropy Production as an Inference Algorithm that Translates Physical Assumptions into Macroscopic Predictions: Don’t Shoot the Messenger

Is Maximum Entropy Production (MEP) a physical principle? In this paper I tentatively suggest it is not, on the basis that MEP is equivalent to Jaynes’ Maximum Entropy (MaxEnt) inference algorithm that passively translates physical assumptions into macroscopic predictions, as applied to non-equilibrium systems. MaxEnt itself has no physical content; disagreement between MaxEnt predictions and experiment falsifies the physical assumptions, not MaxEnt. While it remains to be shown rigorously that MEP is indeed equivalent to MaxEnt for systems arbitrarily far from equilibrium, work in progress tentatively supports this conclusion. In terms of its role within non-equilibrium statistical mechanics, MEP might then be better understood as Messenger of Essential Physics.

Spatial analysis of plague in California: niche modeling predictions of the current distribution and potential response to climate change

BackgroundPlague, caused by the bacterium Yersinia pestis, is a public and wildlife health concern in California and the western United States. This study explores the spatial characteristics of positive plague samples in California and tests Maxent, a machine-learning method that can be used to develop niche-based models from presence-only data, for mapping the potential distribution of plague foci. Maxent models were constructed using geocoded seroprevalence data from surveillance of California ground squirrels (Spermophilus beecheyi) as case points and Worldclim bioclimatic data as predictor variables, and compared and validated using area under the receiver operating curve (AUC) statistics. Additionally, model results were compared to locations of positive and negative coyote (Canis latrans) samples, in order to determine the correlation between Maxent model predictions and areas of plague risk as determined via wild carnivore surveillance.ResultsModels of plague activity in California ground squirrels, based on recent climate conditions, accurately identified case locations (AUC of 0.913 to 0.948) and were significantly correlated with coyote samples. The final models were used to identify potential plague risk areas based on an ensemble of six future climate scenarios. These models suggest that by 2050, climate conditions may reduce plague risk in the southern parts of California and increase risk along the northern coast and Sierras.ConclusionBecause different modeling approaches can yield substantially different results, care should be taken when interpreting future model predictions. Nonetheless, niche modeling can be a useful tool for exploring and mapping the potential response of plague activity to climate change. The final models in this study were used to identify potential plague risk areas based on an ensemble of six future climate scenarios, which can help public managers decide where to allocate surveillance resources. In addition, Maxent model results were significantly correlated with coyote samples, indicating that carnivore surveillance programs will continue to be important for tracking the response of plague to future climate conditions.