Assessing the effect of prevalence on the predictive performance of species distribution models using simulated data

By Truly Santika. 2011

87 citations recorded by the source corpus.

Subfields: Ecological Modeling, Ecology, Nature and Landscape Conservation

View Paper: https://www.semanticscholar.org/paper/10c4328187780ea373810bb87ee67ec03aa17299

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Ecological Modeling

In Ecology

In Nature and Landscape Conservation

Citations in the corpus

Insights into the area under the receiver operating characteristic curve (AUC) as a discrimination measure in species distribution modelling (2012, background): “The proper evaluation of models is a key challenge in SDM (Vaughan & Ormerod, 2005) and despite the wide acceptance of some standard methods/statistics, there is currently a strong ongoing debate as to their use (e.g.Allouche et al., 2006; JiménezValverde et al., 2008; e.g. Lobo et al., 2008; Peters”
Drawbacks to palaeodistribution modelling: the case of South American seasonally dry forests (2013, result): “…the prevalence of species was kept constant in the process of obtaining the pseudo-absences, and in fact this factor affects the SDM algorithms differently (see Santika, 2011); thus, variation in prevalence among species can explain the much higher variance component of SDMs in previous studies.”
Invasive species distribution models – how violating the equilibrium assumption can create new insights (2012, background): “A threshold of 0.5 is not usually optimal when model outputs are true probabilities (Santika, 2011), but given that the outputs were the agreements between models, we selected this value in order to make sure that the majority of model agreements were decided (i.e. at least half of the models…”
Stacking species distribution models and adjusting bias by linking them to macroecological models (2014, background): “A signal of prevalence in model accuracy has been consistently reported, with SDMs for both common and rare species being less reliable than those for species of intermediate prevalence (McPherson et al., 2004; Santika, 2011).”
Improving species distribution model predictive accuracy using species abundance: Application with boosted regression trees (2020, result): “Although species with low prevalence can result in imbalanced data, causing biased predictions in SDMs as has been indicated in a number of studies (Manel et al., 2001; McPherson and Jetz, 2007; Santika, 2011), this issue has seldom been discussed in the context of BRT models.”
Areas of Climate Stability of Species Ranges in the Brazilian Cerrado: Disentangling Uncertainties Through Time (2012, background): “…effects of different sources of uncertainties, including the methodological ones (i.e., from ENMs and AOGCMs) and those from 154 Natureza & Conservação 10(2):152-159, December 2012Terribile et al. across the Neotropical grid to fit the model, keeping prevalence equal to 0.25 (see Santika 2011).”
High uncertainty in the effects of data characteristics on the performance of species distribution models (2021, methodology): “Variations in SDM performance have been attributed to the differences in the spatial resolutions of the analyses, the particular SDM technique used and the evaluation metric used to assess model performance (Hijmans, 2012; Marmion et al., 2009; Santika, 2011; Wogan, 2016).”
Testing the ability of species distribution models to infer variable importance (2019, background): “Adding to these complications, many metrics of model performance are confounded by extent and prevalence (Santika 2011; Smith 2013b; Somodi et al. 2017; Sor et al. 2017), which compromises their use as metrics of variable importance (Figs.”
A comprehensive evaluation of predictive performance of 33 species distribution models at species and community levels (2019, background): “…data (Fielding and Haworth 1995), especially the prevalence of species (Leathwick et al. 2006, Meynard and Quinn 2007, Syphard and Franklin 2009, Santika 2011, Madon et al. 2013) and the strength and shape of the environmental gradient (Thuiller et al. 2003, Austin et al. 2006, Santika and…”
A strict maximum likelihood explanation of MaxEnt, and some implications for distribution modelling (2013, background): “In a recent study, Santika (2011) used simulated data to demonstrate that the relationship between AUC and prevalence is context dependent, influenced by the strength of the relationship between the modelled target and the dominant explanatory variable, the shape of the overall response curve with…”