Selecting pseudo‐absences for species distribution models: how, where and how many?

By Morgane Barbet-Massin, Frédéric Jiguet, Cécile H. Albert, Wilfried Thuiller. 2012

2,534 citations recorded by the source corpus.

Subfields: Ecological Modeling, Ecology, Nature and Landscape Conservation

View Paper: https://www.semanticscholar.org/paper/e3da30929bcaaf6e904fe7756f620ea164e00fc5

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

Protected areas network is not adequate to protect a critically endangered East Africa Chelonian: Modelling distribution of pancake tortoise, Malacochersus tornieri under current and future climates (2020, methodology): “[66], as follows: MaxEnt, a machine-learning approach, with 1,000 randomly selected pseudo-absences; Multivariate Adaptive Regression Splines, a regression-based approach, with 100 randomly selected pseudo-absences; Multiple Discriminant Analysis, a classification approach, with 100 pseudo-absences ”
A 40‐year, continent‐wide, multispecies assessment of relevant climate predictors for species distribution modelling (2014, background): “…given that confirmed absences are usually difficult to obtain, especially for mobile species (Mackenzie & Royle, 2005), pseudo-absences are widely used, adding yet another source of uncertainty (Chefaoui & Lobo, 2008; VanDerWal et al., 2009; Wisz & Guisan, 2009; Barbet-Massin et al., 2012a).”
Identifying predictable foraging habitats for a wide‐ranging marine predator using ensemble ecological niche models (2015, methodology): “Five iterations of 1000 randomly selected control locations were used over successive model runs (Barbet-Massin, 2012).”
The global distribution and burden of dengue (2013, neutral): “Asia bore 70% (67 [47-94] million infections) of this burden, and is characterised by large swathes of densely populated regions coinciding with very high suitability for disease transmission.”
Present and future distribution of three aquatic plants taxa across the world: decrease in native and increase in invasive ranges (2017, methodology): “These algorithms require datasets with both presence and absence, and we had presenceonly records, so we randomly generated three sets of pseudo-absences with equal numbers of presences (Barbet-Massin et al. 2012).”
A framework for species distribution modelling with improved pseudo-absence generation (2015, methodology): “%) via the MAXENT algorithm (see e.g.: Barbet-Massin et al., 2012;39 Jiménez-Valverde, 2012, for details), other 28% used profile techniques and40 a 12% used target group background (note that some of the articles anal-41 ysed used more than one type of technique, and therefore percentages do42…”
Predicting the distribution of vulnerable marine ecosystems in the deep sea using presence-background models (2014, background): “A number of studies have observed significant changes in model behaviour as a result of altering the quantity and location of the background data used to fit SDMs (Barbet-Massin et al., 2012; Lobo and Tognelli, 2011; Phillips et al., 2009).”
Accounting for spatially biased sampling effort in presence‐only species distribution modelling (2015, methodology): “Random background points [n = 10,000, as recommended by Barbet-Massin et al. (2012)] were generated using the Geospatial Modelling Environment (GME) (Beyer, 2012).”
Bioclimatic velocity: the pace of species exposure to climate change (2014, methodology): “We randomly selected 5000 absences from the many absences available for each species in the presence/absence dataset, allowing for a large number of absences covering the environmental space (Barbet-Massin et al., 2012).”
On the selection of thresholds for predicting species occurrence with presence‐only data (2015, result): “This strategy was adopted by Liu et al. (2013a, 2013b), and similar strategy has already been recommended by others (Barbet-Massin et al. 2012).”
Integrating ensemble species distribution modelling and statistical phylogeography to inform projections of climate change impacts on species distributions (2013, background): “Background records were weighted to maintain equal prevalence; this has been found to maximize model performance (Barbet-Massin et al., 2012).”
SDMtoolbox: a python‐based GIS toolkit for landscape genetic, biogeographic and species distribution model analyses (2014, background): “Thus, a larger study spatial extent can lead to the selection of a higher proportion of less informative background points (Barbet-Massin et al. 2012).”
Practical guidance on characterizing availability in resource selection functions under a use-availability design. (2013, background): “In addition, Barbet-Massin et al. (2012) suggest that the modeling framework (e.g., GLM, GAM, or machine learning methods) can influence the number of availability points needed.”