Rethinking receiver operating characteristic analysis applications in ecological niche modeling

Peterson, A. Townsend; Papeş, Monica; Soberón, Jorge

doi:10.1016/j.ecolmodel.2007.11.008

Cited by 1,391 publications

(1,211 citation statements)

References 35 publications

Supporting

Mentioning

1,141

Contrasting

Unclassified

Order By: Relevance

“…They agree these AUC scores indicate that predictions based on presence-only data can be sufficiently accurate to be used in conservation planning and in numerous other applications in which estimates of species distribution are relevant. Many discussions had permeated the misleading use of AUC values as an indicator of model performance (e.g., Lobo et al 2008, Peterson et al 2008. Despite some of the limitations imposed on the performance of ROC analysis, it has been widely used to assess the accuracy of predictive distribution models.…”

Section: Discussionmentioning

confidence: 99%

Assessment of Cerdocyon thous distribution in an agricultural mosaic, southeastern Brazil

Ferraz

Siqueira²,

Martin³

et al. 2010

Mamm

View full text Add to dashboard Cite

Crab-eating fox, Cerdocyon thous, is a habitat generalist species relatively common throughout its range inhabiting most environments owing to its low habitat requirements. Because no information is available for anthropogenic habitats, this study aimed to describe the species occurrence in a highly fragmented and heterogeneous landscape in southeastern Brazil. C. thous was surveyed in 95 study sites in four main land covers (native forest, eucalyptus forest, sugar cane and pasture) from April to September 2006. Presence records (ns28) and landscape variables (land cover, heterogeneity, stream and forest fragment distance, elevation and slope) were used for modeling in Maxent. The bootstrapping method was used for sampling 70% of the dataset for training and 30% for testing models. The species was equally distributed in all types of land cover, although it was more frequent in the sugar cane areas and more associated with forest fragments and heterogeneous habitats. The potential distribution model predicted forest patches and its surroundings as highly suitable for the species. It also predicted part of the sugar cane matrix as highly suitable, probably related to prey availability. Results suggested that the anthropogenic landscape studied encompasses many suitable habitats for species occurrence, resulting in the necessity to assess its potential role in vertebrate communities.

show abstract

Section: Discussionmentioning

confidence: 99%

Assessment of Cerdocyon thous distribution in an agricultural mosaic, southeastern Brazil

Ferraz

Siqueira²,

Martin³

et al. 2010

Mamm

View full text Add to dashboard Cite

show abstract

“…We set up parameters for a Maxent model according to a MaxEnt Model v3.3.3e Tutorial (Young et al, 2011). The accuracy of each model prediction for each taxon was quantified by calculating the area under the 'receiver operating characteristic (ROC) curve' (AUC) (Peterson et al, 2008;Elith and Leathwick, 2009). AUC values above 0Á7 are considered to indicate good model performance (Fielding and Bell, 1997).…”

Section: Ecological Niche Modellingmentioning

confidence: 99%

Molecular phylogeny, biogeography and ecological niche modelling ofCardiocrinum(Liliaceae): insights into the evolutionary history of endemic genera distributed across the Sino-Japanese floristic region

Yang

Xie

et al. 2016

Ann Bot

View full text Add to dashboard Cite

Background and Aims The patterns of evolutionary assembly in the Sino-Japanese floristic region (SJFR) remain largely unknown due to a lack of integrative multidimensional studies throughout the region. To address this issue, we elucidated the evolutionary history of Cardiocrinum (Liliaceae), a genus containing four taxa distributed across the SJFR.Methods Fifty-four populations were sampled throughout the geographical range of Cardiocrinum to assess genetic structure, analyse phylogenetic relationships and reconstruct ancestral area based on six chloroplast DNA (cpDNA) fragments and three low copy nuclear genes (LCNG). Ecological niche modelling was used to examine the potential range shifts of Cardiocrinum in response to climatic change.Key Results The molecular data showed high genetic similarity in the cpDNA (98Á37 %) and LCNG (94Á53 %) sequences. The biogeographical analyses revealed that the ancestor of Cardiocrinum diversified during the late Miocene (approx. 7Á32 Mya) in Central China. The ancestor of the C. giganteum lineage dispersed westward to the Himalayas and south-west China with the split between C. giganteum and C. giganteum var. yunnanense occurring around 4Á11 Mya consistent with the period of orogeny of the Hengduan Mountains. Some populations of the C. cathayanum lineage dispersed eastward to south Japan via the land bridge approx. 4Á97 Mya, providing opportunities for allopatric speciation of C. cordatum. The predicted suitable habitats of Cardiocrinum have become smaller and more fragmented since the Last Glacial Maximum.Conclusions Our study provides evidence of a biogeographical pattern of dispersal from Central China to the Himalayas in the west and Japan in the east for genera distributed across the SJFR, and highlights that the orogeny of the Hengduan Mountains and fluctuations of the sea level of the East China Sea played important roles in promoting species divergence.

show abstract

“…Comparative studies can provide guidelines for using the methods in the current time, but the problem is that they cannot be understood as final solutions to the problem if there are conflicting results, as we show here when comparing GARP and MAXENT (see also Peterson et al, 2008). First of all, the basic idea of accumulating results from multiple studies is that they must be directly comparable, both in terms of ENMs used and their evaluation criterion.…”

Section: Discussionmentioning

confidence: 99%

“…Also, as recently pointed out by Peterson et al (2008), the tests by Elith et al (2006) were actually performed using high-quality data and designed to evaluate ENMs in a situation of fine-scale modeling of species distribution details. However, in some circumstances it is necessary to project distributions into large and unknown regions in which samples are sparse or non-existent (i.e., transferability).…”

Section: Discussionmentioning

confidence: 99%

“…The AUC varies from 0 to 1, where a score of 1 indicates perfect discrimination between areas where a species is present, versus those where it is absent. Although AUC have recently been criticised (see Lobo et al, 2008;Peterson et al, 2008), it can provide at least a preliminary indication of the usefulness of the distribution models for identifying suitable areas of occurrence for particular species (Elith et al, 2006). We also recorded, for each species, the area of occurrence predicted by GARP and MAXENT, using the threshold generated to "cut" the potential distribution.…”

Section: Datamentioning

confidence: 99%

See 1 more Smart Citation

How many studies are necessary to compare niche-based models for geographic distributions? Inductive reasoning may fail at the end

2010

View full text Add to dashboard Cite

The use of ecological niche models (ENM) to generate potential geographic distributions of species has rapidly increased in ecology, conservation and evolutionary biology. Many methods are available and the most used are Maximum Entropy Method (MAXENT) and the Genetic Algorithm for Rule Set Production (GARP). Recent studies have shown that MAXENT perform better than GARP. Here we used the statistics methods of ROC -AUC (area under the Receiver Operating Characteristics curve) and bootstrap to evaluate the performance of GARP and MAXENT in generate potential distribution models for 39 species of New World coral snakes. We found that values of AUC for GARP ranged from 0.923 to 0.999, whereas those for MAXENT ranged from 0.877 to 0.999. On the whole, the differences in AUC were very small, but for 10 species GARP outperformed MAXENT. Means and standard deviations for 100 bootstrapped samples with sample sizes ranging from 3 to 30 species did not show any trends towards deviations from a zero difference in AUC values of GARP minus AUC values of MAXENT. Ours results suggest that further studies are still necessary to establish under which circumstances the statistical performance of the methods vary. However, it is also important to consider the possibility that this empirical inductive reasoning may fail in the end, because we almost certainly could not establish all potential scenarios generating variation in the relative performance of models.Keywords: coral snakes, Elapidae, potential distribution models, GARP, MAXENT.Quantos estudos ainda serão necessários para avaliar modelos de distribuição geográfica baseados em modelagem do nicho?Um pensamento indutivo pode estar fadado ao fracasso

show abstract

Rethinking receiver operating characteristic analysis applications in ecological niche modeling

Cited by 1,391 publications

References 35 publications

Assessment of Cerdocyon thous distribution in an agricultural mosaic, southeastern Brazil

Assessment of Cerdocyon thous distribution in an agricultural mosaic, southeastern Brazil

Molecular phylogeny, biogeography and ecological niche modelling ofCardiocrinum(Liliaceae): insights into the evolutionary history of endemic genera distributed across the Sino-Japanese floristic region

How many studies are necessary to compare niche-based models for geographic distributions? Inductive reasoning may fail at the end

Contact Info

Product

Resources

About